Method, apparatus and system for wireless vital monitoring using radio frequency signals

Abstract

Methods, apparatuses, and systems for wireless monitoring systems, wireless vital monitoring, and wireless proximity sensing are described.SOLUTION: A system includes a transmitter configured to transmit a first wireless signal over a wireless channel of a venue, a receiver configured to receive a second wireless signal over the wireless channel, and a processor. The second wireless signal comprises a reflection of the first wireless signal by at least one living being having at least one repetitive motion at the location. The processor is configured to obtain a time series (TSCI) of channel information of the wireless channel based on the second wireless signal, to generate, for each living being of the at least one living being, a vital signal representative of all repetitive movements of the living being based on the TSCI, to extract a heart rate signal from the vital signal of each living being, and to monitor heart rate variability based on the heart rate signal for each living being at the location.SELECTED DRAWING: Figure 1A

Description

[Technical Field]

[0001] The present disclosure relates generally to wireless motion tracking, wireless vitals monitoring, and wireless proximity sensing. More specifically, the present disclosure relates to wireless tracking of micro-motions such as keystrokes on a surface by processing wireless channel information (CI), heart rate tracking and monitoring by processing wireless CI, and motion detection by object proximity information by processing wireless CI. [Background technology]

[0002] As the primary and most integrated computer peripheral, keyboards have become indispensable in our daily lives. However, physical keyboards have suffered from portability issues. Furthermore, as IoT (Internet of Things) devices become smaller, they generally cannot afford to have bulky physical keyboards. Therefore, virtual keyboards have been highly demanded as a convenient replacement for regular physical keyboards or to enable keyboard-less typing experiences on billions of IoT devices.

[0003] Virtual keyboards can be implemented either actively or passively. Active approaches require wearable sensors attached to the user, which is inconvenient, encouraging the design of passive, contactless systems, such as vision-based approaches that utilize visual technologies based on the output from cameras, lasers or infrared, acoustic-based, and electromagnetic radiation-based approaches. However, vision-based approaches raise concerns about privacy intrusion and are sensitive to lighting conditions, while acoustic-based approaches suffer from false alarms due to ambient interference.

[0004] Heart rate variability (HRV), defined as the variation in the period between successive heartbeats, i.e., inter-beat intervals (IBI), is an important indicator of an individual's overall health. Analysis of HRV has proven to be a powerful tool for assessing cardiac health and the state of the autonomic nervous system. High-precision HRV monitoring is required in many applications, such as early diagnosis of cardiovascular diseases, stress assessment, emotion recognition, and anxiety treatment.

[0005] Traditional measurements of HRV are obtained by continuously measuring IBI using electrocardiogram (ECG) or photoplethysmography (PPG) sensors, both of which are dedicated medical devices that must be in physical contact with human skin. However, using ECG or PPG can be uncomfortable for users and can sometimes cause skin allergies. To avoid direct contact with the user's skin, other wearable devices, such as inertial measurement units (IMUs), have been explored to measure chest surface motion to determine IBI and subsequently measure HRV. While some of the aforementioned methods are less invasive than ECG- and PPG-based approaches, all of them require the user to wear dedicated devices, which are cumbersome and usually expensive for routine use. Therefore, it is desirable to monitor HRV in a non-contact and accurate manner with a robust system.

[0006] Indoor motion detection plays a major role in modern security systems, smart homes, and healthcare. However, common approaches relying on video, infrared, ultra-wideband (UWB), etc., require specialized hardware deployment and have their own limitations, such as good visibility requirements, line-of-sight (LOS) constraints, and privacy issues.

[0007] In recent years, Wi-Fi has expanded its role from a communication medium to a wireless sensing tool due to its ubiquitous deployment and cost-effectiveness. Existing approaches, particularly those based on fine-grained physical layer channel state information (CSI), have investigated the feasibility of using Wi-Fi signals to detect the presence of motion. In real-world applications, it is not just the presence of motion that matters, but also the distance at which the motion occurs. For example, smart home technologies can fully realize potential energy savings for the automatic control of utilities such as lighting, heating, and indoor surveillance cameras only if motion is detected within a small range. Furthermore, location-aware motion detection can provide important location-related context for further activity recognition in daily life. However, most existing approaches only aim to detect motion within a predefined (usually large) area without examining how far the target motion is. Therefore, they are not suitable for recognizing motion in close proximity. Furthermore, reliable motion proximity detectors are still needed to enable smart devices and systems to work efficiently by sensing nearby moving objects.

[0008] Previous approaches have also attempted to localize human targets using fingerprinting-based or geometric mapping-based methods. However, they typically require high setup effort, such as multiple transceivers with specific geometric arrangements or dedicated calibration, and training before achieving even accurate localization, and are therefore impractical for motion proximity sensing. Because proximity sensing does not require highly accurate positioning, the high complexity of the motion localization approaches described above is not necessary for proximity sensing. Summary of the Invention

[0009] The present disclosure generally relates to wireless motion tracking, wireless vitals monitoring, and wireless proximity sensing. More specifically, the present disclosure relates to wireless tracking of micro-motions such as keystrokes by processing wireless channel information (CI), heart rate tracking and monitoring by processing wireless CI, and motion detection using object proximity information by processing wireless CI.

[0010] In one embodiment, a system for wirelessly tracking keystrokes on a surface is described, the system including: a transmitter configured to transmit a first wireless signal over a wireless channel of a location including the surface using a transmitting antenna; a receiver configured to receive a second wireless signal over the wireless channel using a plurality of receiving antennas, the second wireless signal comprising a reflection of the first wireless signal by at least one moving object in the location; and a processor. The processor is configured to: acquire, for each of the plurality of receive antennas, a time series of channel information (CI) of the wireless channel based on the second wireless signal, each CI including at least one of channel state information (CSI), channel impulse response (CIR), channel frequency response (CFR), or received signal strength index (RSSI); detect at least one keystroke on the surface based on the time series of CI (TSCI) obtained for each of the plurality of receive antennas; determine at least one position of the at least one keystroke on the surface; and determine at least one key associated with the at least one keystroke based on the at least one position.

[0011] In another embodiment, a wireless device of a wireless tracking system is described. The wireless device includes a processor, a memory communicatively coupled to the processor, and a receiver communicatively coupled to the processor. An additional wireless device of the wireless tracking system is configured to transmit a first wireless signal over a wireless channel of a location including a surface. The receiver is configured to receive a second wireless signal over the wireless channel. The second wireless signal includes reflections of the first wireless signal by at least one moving object in the location. The processor is configured to: acquire a time series of channel information (CI) for the wireless channel based on the second wireless signal, each CI including at least one of channel state information (CSI), channel impulse response (CIR), channel frequency response (CFR), or received signal strength index (RSSI); detect at least one keystroke on the surface based on the time series of CI (TSCI); determine at least one location of the at least one keystroke on the surface; and determine at least one key associated with the at least one keystroke based on the at least one location.

[0012] In yet another embodiment, a method for a wireless tracking system is described. The method includes: transmitting a first wireless signal through a wireless channel at a location including a surface; receiving a second wireless signal through the wireless channel, the second wireless signal including a reflection of the first wireless signal by at least one moving object at the location; acquiring a time series of channel information (CI) for the wireless channel based on the second wireless signal, each CI including at least one of channel state information (CSI), channel impulse response (CIR), channel frequency response (CFR), or received signal strength index (RSSI); detecting at least one keystroke on the surface based on the time series of CI (TSCI), determining at least one location of the at least one keystroke on the surface, and determining at least one key associated with the at least one keystroke based on the at least one location.

[0013] In one embodiment, a system for wireless monitoring is described. The system includes a transmitter configured to transmit a first wireless signal over a wireless channel at a location using N1 transmit antennas, a receiver configured to receive a second wireless signal over the wireless channel using N2 receive antennas, where N1 and N2 are positive integers. The second wireless signal includes reflections of the first wireless signal by at least one living thing having at least one repetitive movement at the location. The processor is configured to: acquire a plurality of time series of channel information (TSCIs) for the wireless channel based on the second wireless signal, each of the plurality of TSCIs associated with a respective transmit antenna of the transmitter and a respective receive antenna of the receiver; generate a vital signal for each of the at least one living thing representing all repetitive movements of the living thing based on the plurality of TSCIs; extract a heart rate signal from the vital signal for each of the living thing; and monitor heart rate variability for each living thing at the location based on the heart rate signal.

[0014] In another embodiment, a wireless device of a wireless monitoring system is described. The wireless device includes a processor, a memory communicatively coupled to the processor, and a receiver communicatively coupled to the processor. An additional wireless device of the wireless monitoring system is configured to transmit a first wireless signal over a wireless channel at a location. The receiver is configured to receive a second wireless signal over the wireless channel. The second wireless signal includes a reflection of the first wireless signal by at least one living thing having at least one repetitive movement at the location. The processor is configured to: obtain a time series of channel information (TSCI) for the wireless channel based on the second wireless signal; generate, for each of the at least one living thing, a vital signal representing all of the repetitive movements of the living thing based on the TSCI; extract a heart rate signal from the vital signal of each living thing; and monitor heart rate variability for each living thing at the location based on the heart rate signal.

[0015] In yet another embodiment, a method for a wireless monitoring system is described, the method including: transmitting a first wireless signal through a wireless channel at a location; receiving a second wireless signal through the wireless channel, the second wireless signal including reflections of the first wireless signal by a plurality of people at the location; acquiring a time series of channel information (TSCI) for the wireless channel based on the second wireless signal, each CI including at least one of a channel state information (CSI), a channel impulse response (CIR), a channel frequency response (CFR), or a received signal strength index (RSSI); generating, for each of the plurality of people, a vital signal representing all repetitive movements of the person based on the TSCI; extracting a heart rate signal from the vital signal for each person; and simultaneously monitoring heart rate variability for each of the plurality of people based on the heart rate signal.

[0016] In one embodiment, a system for wireless proximity sensing is described. The system includes a transmitter configured to transmit a first wireless signal over a wireless multipath channel of a location, a receiver configured to receive a second wireless signal over the wireless multipath channel, and a processor. The second wireless signal differs from the first wireless signal due to the wireless multipath channel being affected by movement of an object at the location. The processor is configured to: obtain a time series of channel information (TSCI) of the wireless multipath channel based on the second wireless signal, where each TSCI channel information (CI) includes a plurality of CI components, each CI associated with an index; calculate inter-component statistics based on the plurality of CI components; calculate proximity information of the object relative to a reference position within the location based on the inter-component statistics; and perform a task based on the proximity information of the object.

[0017] In another embodiment, a wireless device of a wireless proximity sensing system is described. The wireless device includes a processor, a memory communicatively coupled to the processor, and a receiver communicatively coupled to the processor. An additional wireless device of the wireless proximity sensing system is configured to transmit a first wireless signal over a wireless multipath channel of a location. The receiver is configured to receive a second wireless signal over the wireless multipath channel. The second wireless signal differs from the first wireless signal due to the wireless multipath channel being affected by movement of an object at the location. The processor is configured to: obtain a time series of channel information (TSCI) of the wireless multipath channel based on the second wireless signal, where each TSCI channel information (CI) includes multiple CI components, each CI component associated with an index; calculate inter-component statistics based on the multiple CI components; calculate proximity information of the object relative to a reference position within the location based on the inter-component statistics; and perform a task based on the proximity information of the object.

[0018] In yet another embodiment, a wireless proximity sensing system is described, the method being configured to: transmit a first wireless signal through a wireless multipath channel at a location; receive a second wireless signal through the wireless multipath channel, the second wireless signal being different from the first wireless signal due to the wireless multipath channel being affected by a movement of an object within the location; acquire a time series of channel information (TSCI) of the wireless multipath channel based on the second wireless signal, where each channel information (CI) of the TSCI includes multiple CI components; calculate inter-component statistics based on the multiple CI components; calculate proximity information of the object relative to a reference position within the location based on the inter-component statistics; and perform a task based on the proximity information of the object.

[0019] In yet another embodiment, a wireless proximity sensing system is described, the method including: asynchronously transmitting a plurality of wireless signals from one of a plurality of Type 1 heterogeneous wireless devices of the wireless proximity sensing system located at various positions in a location through a wireless multipath channel of the location, asynchronously receiving the plurality of wireless signals by a Type 2 heterogeneous wireless device of the wireless proximity sensing system through the wireless multipath channel, where each received wireless signal differs from the respective transmitted wireless signal due to the wireless multipath channel being affected by a movement of an object within the location; obtaining a plurality of time series of channel information (TSCI) of the wireless multipath channel based on the respective received wireless signals, where each channel information (CI) of each TSCI includes a respective plurality of CI components; calculating a plurality of individual inter-component statistics, each individually calculated based on the CI components of the respective TSCI; calculating joint inter-component statistics based on the CI components associated with each TSCI; and calculating proximity information of the object based on the joint inter-component statistics or at least one of the plurality of individual inter-component statistics.

[0020] Other concepts relate to software for implementing the present disclosure relating to wireless motion tracking, wireless vital signs monitoring, and wireless proximity sensing. Additional novel features will be set forth in part in the description which follows, and in part will become apparent to those skilled in the art upon examination of the following drawings and accompanying drawings, or may be learned by the production or operation of the embodiments. The novel features of the present disclosure may be realized and attained by practice or use of various aspects of the methods, instrumentalities, and combinations that are described in the detailed embodiments set forth below. [Brief explanation of the drawings]

[0021] The methods, systems, and / or devices described herein will be further explained with reference to exemplary embodiments. These exemplary embodiments will be described in detail with reference to the drawings. These embodiments are non-limiting exemplary embodiments, and like reference numerals represent like structure throughout the several views of the drawings.

[0022] [Figure 1A] 1 illustrates an exemplary device configuration for a virtual keyboard system, according to some embodiments of the present disclosure.

[0023] [Figure 1B] 1 illustrates an exemplary coordinate system for a virtual keyboard system, according to some embodiments of the present disclosure.

[0024] [Figure 1C] 1 illustrates an example frame structure illustrating the concept of burst and pulse radar in millimeter wave (mmWave) radio, according to some embodiments of the present disclosure.

[0025] [Figure 2] 1 illustrates an exemplary workflow for wirelessly tracking keystrokes according to some embodiments of the present disclosure.

[0026] [Figure 3A] 1 illustrates an example channel impulse response (CIR) amplitude difference for a stationary reference frame in accordance with some embodiments of the present disclosure.

[0027] [Figure 3B] 10 illustrates an exemplary quantile-quantile (QQ) plot of CIR amplitude differences of reference frames, according to some embodiments of the present disclosure.

[0028] [Figure 3C] 10 illustrates an example CIR amplitude difference for a frame containing a keystroke, according to some embodiments of the present disclosure.

[0029] [Figure 4A] , [Figure 4B] , [Figure 4C] , [Figure 4D] 1 illustrates features for motion identification according to some embodiments of the present disclosure.

[0030] [Figure 5A] 1 illustrates mixed signals from an object and background during keystroke tracking, according to some embodiments of the present disclosure.

[0031] [Figure 5B] 1 illustrates an exemplary adaptive background cancellation according to some embodiments of the present disclosure.

[0032] [Figure 6] 1 illustrates exemplary spatial spectra of 1-key, 2-key, and 3-key keystrokes according to some embodiments of the present disclosure.

[0033] [Figure 7A] , [Figure 7B] , [Figure 7C] 1 illustrates an exemplary spatial spectrum according to some embodiments of the present disclosure.

[0034] [Figure 8A] , [Figure 8B] 1 illustrates an exemplary geometric model for keyboard calibration, according to some embodiments of the present disclosure.

[0035] [Figure 9] 1 illustrates a generalized two-dimensional case for keyboard calibration, according to some embodiments of the present disclosure.

[0036] [Figure 10A] , [Figure 10B] 1 illustrates exemplary performance on a virtual computer keyboard according to some embodiments of the present disclosure.

[0037] [Figure 11A] , [Figure 11B] 1 illustrates an exemplary performance on a virtual piano keyboard according to some embodiments of the present disclosure.

[0038] [Figure 12A] , [Figure 12B] 1 illustrates an exemplary multi-keystroke precision according to some embodiments of the present disclosure.

[0039] [Figure 13] 1 illustrates a flowchart of an exemplary method for wirelessly tracking keystrokes, according to some embodiments of the present disclosure.

[0040] [Figure 14A] 1 illustrates an exemplary setup for a wireless vital signs monitoring system, according to some embodiments of the present disclosure.

[0041] [Figure 14B] 1 illustrates an exemplary workflow for wirelessly monitoring heart rate variability according to some embodiments of the present disclosure.

[0042] [Figure 15] 1 illustrates an exemplary basic concept of a frequency modulated continuous wave (FMCW) radar system according to some embodiments of the present disclosure.

[0043] [Figure 16] 1 illustrates an exemplary antenna deployment for a wireless vital signs monitoring system, according to some embodiments of the present disclosure.

[0044] [Figure 17A] , [Figure 17B] , [Figure 17C] , [Figure 17D]1 illustrates exemplary performance of a reflective object detector according to some embodiments of the present disclosure.

[0045] [Figure 18A] , [Figure 18B] , [Figure 18C] , [Figure 18D] 1 illustrates exemplary performance of a subject detector according to some embodiments of the present disclosure.

[0046] [Figure 19A] , [Figure 19B] 10 illustrates exemplary performance of a heart rate extractor according to some embodiments of the present disclosure.

[0047] [Figure 20A] , [Figure 20B] , [Figure 20C] 1 illustrates an exemplary interbeat interval (IBI) estimation according to some embodiments of the present disclosure.

[0048] [Figure 21A] , [Figure 21B] 10 illustrates an example of an IBI estimation error, according to some embodiments of the present disclosure.

[0049] [Figure 22] 1 illustrates a flowchart of an exemplary method for wireless vital monitoring, according to some embodiments of the present disclosure.

[0050] [Figure 23] 1 illustrates an exemplary configuration for a wireless proximity sensing system according to some embodiments of the present disclosure.

[0051] [Figure 24A] , [Figure 24B] 1 illustrates an exemplary power delay profile in accordance with some embodiments of the present disclosure.

[0052] [Figure 25A] 10 illustrates an example distribution of the phase sum of the dynamic part of the channel state information across adjacent subcarriers, in accordance with some embodiments of the present disclosure.

[0053] [Figure 25B] 10 illustrates an example distribution of phase differences for a static portion of CSI, in accordance with some embodiments of the present disclosure.

[0054] [Figure 26] 10 illustrates an example CSI power response over time for two adjacent subcarriers in a vacant case, in accordance with some embodiments of the present disclosure.

[0055] [Figure 27] 10 illustrates an example CSI power response over time for two adjacent subcarriers when motion is closer to the receiver, in accordance with some embodiments of the present disclosure.

[0056] [Figure 28A] , [Figure 28B] , [Figure 28C] 10A-10C illustrate example correlation matrices for a person moving at distances of 1 m, 3 m, and 5 m, respectively, according to some embodiments of the present disclosure.

[0057] [Figure 29A] , [Figure 29B] , [Figure 29C] 1 shows an example covariance matrix when a person moves at distances of 1 m, 3 m, and 5 m, according to some embodiments of the present disclosure.

[0058] [Figure 30] 1 illustrates a flowchart of an exemplary method for wireless proximity sensing, according to some embodiments of the present disclosure.

[0059] [Figure 31]1 illustrates an example block diagram of a first wireless device of a wireless system, in accordance with some embodiments of the present disclosure.

[0060] [Figure 32] 1 illustrates an example block diagram of a second wireless device of a wireless system, in accordance with some embodiments of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0061] In the following detailed description, numerous specific details are set forth, by way of example, to provide a thorough understanding of the relevant disclosure. However, it should be apparent to those skilled in the art that the present disclosure may be practiced without such details. In other instances, well-known methods, procedures, components, and / or circuits have been described at a relatively high level, without detail, in order to avoid unnecessarily obscuring aspects of the present disclosure.

[0062] In one embodiment, the present disclosure discloses a method, apparatus, device, system, and / or software (method / apparatus / device / system / software) for a wireless monitoring system. Time-series channel information (CI) of a wireless multipath channel may be obtained (e.g., dynamically) using a processor, a memory communicatively coupled to the processor, and a set of instructions stored in the memory. A time series of CI (TSCI) may be extracted from wireless signals (signals) transmitted between a type 1 heterogeneous wireless device (e.g., a wireless transmitter, TX) and a type 2 heterogeneous wireless device (e.g., a wireless receiver, RX) at a venue over the channel. The channel may be affected by representations (e.g., motion, movement, representation, and / or position / pose / shape / representation changes) of objects at the venue. Properties and / or spatio-temporal information (STI, e.g., motion information) of objects and / or object motion may be monitored based on the TSCI. A task may be performed based on the properties and / or STI. A presentation associated with the task may be generated in a user interface (UI) on a user's device. The TSCI may be a wireless signal stream. The TSCI or each CI may be pre-processed. The device may be a station (STA). The symbol "A / B" means "A and / or B" in this disclosure.

[0063] Expressions can include placement, placement of movable parts, location, position, orientation, identifiable location, area, spatial coordinates, presentation, state, expression, static representation, size, length, width, height, angle, scale, shape, curve, surface, area, volume, pose, posture, sign, body representation, dynamic representation, dynamic noun, movement, movement sequence, gesture, stretch, contraction, distortion, deformation, body representation (e.g., head, face, eyes, mouth, tongue, hair, voice, neck, limbs, arms, hands, legs, feet, muscles, moving parts), surface representation (e.g., shape, texture, material, color, electromagnetic (EM) properties, visual pattern, humidity, reflectivity, translucency, flexibility), material properties (e.g., living tissue, hair, fabric, metal, wood, leather, plastic, metallic, artificial material, solid, liquid, gas, temperature), movement, activity, behavior, change in expression, and / or any combination.

[0064] Wireless signals include transmit / receive signals, EM radiation, RF signals / transmissions, signals in licensed / unlicensed / ISM bands, band-limited signals, baseband signals, wireless / mobile / cellular communication signals, wireless / mobile / cellular network signals, mesh signals, optical signals / communications, downlink / uplink signals, unicast / multicast / broadcast signals, standard (e.g. WLAN, WWAN, WPAN, WBAN, international, national, industry, de facto, IEEE, IEEE 802, 802.11 / 15 / 16, WiFi, 802.11n / ac / ax / be, 3G / 4G / LTE / 5G / 6G / 7G / 8G, 3GPP, blue tooth, BLE, Zigbee, RFID, UWB, WiMax) compliant signals, protocol signals, standard frames, beacon / pilot / search / inquiry / acknowledge / handshake / synchronization signals, management / control / data signals, standardized wireless / cellular communication protocols, reference signals, source signals, motion probe / detection / The CI may include a sensing signal and / or a series of signals. The wireless signal may include line-of-sight (LOS) and / or non-LOS components (or paths / links). Each CI may be extracted / generated / calculated / detected at a layer of the Type 2 device (e.g., the PHY / MAC layer of the OSI model) and obtained by an application (e.g., software, firmware, driver, app, wireless monitoring software / system).

[0065] A wireless multipath channel can include: a communication channel, an analog frequency channel (e.g., analog carrier frequencies around 700 / 800 / 900 MHz, 1.8 / 1.8 / 2.4 / 3 / 5 / 6 / 27 / 60 GHz), a coded channel (e.g., CDMA), and / or a channel of a wireless network / system (e.g., WLAN, WiFi, mesh, LTE, 4G / 5G, Bluetooth, Zigbee, UWB, RFID, microwave). It can include two or more channels. The channels can be contiguous (e.g., adjacent / overlapping bands) or non-contiguous (e.g., non-overlapping WiFi channels, one at 2.4 GHz and one at 5 GHz).

[0066] The TSCI can be extracted from a wireless signal at a layer of a Type-2 device (e.g., a layer of the OSI reference model, a physical layer, a data link layer, a logical link control layer, a media access control (MAC) layer, a network layer, a transport layer, a session layer, a presentation layer, an application layer, a TCP / IP layer, an Internet layer, or a link layer). The TSCI may also be extracted from a derived signal (e.g., a baseband signal, a motion detection signal, or a motion sensing signal) derived from a wireless signal (e.g., an RF signal). It may be a (wireless) measurement detected by a communication protocol (e.g., a standardized protocol) using an existing mechanism (e.g., a wireless / cellular communication standard / network, 3G / LTE / 4G / 5G / 6G / 7G / 8G, WiFi, IEEE 802.11 / 15 / 16). The derived signal may include a packet having at least one of a preamble, a header, and a payload (e.g., for data / control / management in a wireless link / network). The TSCI may be extracted from a probe signal in a packet (e.g., a training sequence, STF, LTF, L-STF, L-LTF, L-SIG, HE-STF, HE-LTF, HE-SIG-A, HE-SIG-B, or CEF). The motion detection / sensing signal may be recognized / identified based on the probe signal. The packet may be a standard-compliance protocol frame, a management frame, a control frame, a data frame, a sounding frame, an excitation frame, an illumination frame, a null data frame, a beacon frame, a pilot frame, a probe frame, a request frame, a response frame, an association frame, a reassociation frame, a disassociation frame, an authentication frame, an action frame, a report frame, a poll frame, an announcement frame, an extension frame, an inquiry frame, an acknowledgement frame, an RTS frame, a CTS frame, a QoS frame, a CF-Poll frame, a CF-Ack frame, a block acknowledgement frame, a reference frame, a training frame, and / or a synchronization frame.

[0067] The packet may contain control data and / or motion detection probes. Data (e.g., Type 1 device IDs, parameters, characteristics, settings, control signals, commands, instructions, notifications, and broadcast-related information) may be obtained from the payload. The wireless signal may be transmitted by a Type 1 device and received by a Type 2 device. A database (e.g., in a local server, hub device, cloud server, or storage network) may be used to store TSCI, characteristics, STI, signatures, patterns, behaviors, trends, parameters, analysis, output responses, identification information, user information, device information, channel information, location (e.g., map, environment model, network, proximity device / network) information, task information, class / category information, presentation (e.g., UI) information, and / or other information.

[0068] A Type 1 / Type 2 device may include at least one of electronics, circuitry, transmitter (TX) / receiver (RX) / transceiver, RF interface, "origin satellite" / "tracker bot," unicast / multicast / broadcast device, wireless power device, power / destination device, wireless node, hub device, target device, motion detection device, sensor device, remote / wireless sensor device, wireless communication device, wireless enabled device, standard compliant device, and / or receiver. A Type 1 (or Type 2) device may be heterogeneous because, if multiple instances of a Type 1 (or Type 2) device exist, they may have different circuitry, enclosure, structure, purpose, auxiliary functionality, chips / ICs, processor, memory, software, firmware, network connectivity, antenna, brand, model, appearance, form, shape, color, material, and / or specifications. A Type 1 / Type 2 device may include an access point, a router, a mesh router, an Internet of Things (IoT) device, a wireless terminal, one or more radio / RF subsystems / radio interfaces (e.g., 2.4 GHz radio, 5 GHz radio, fronthaul radio, backhaul radio), a modem, an RF front end, an RF / radio chip or integrated circuit (IC).

[0069] At least one of Type 1 devices, Type 2 devices, links between them, objects, characteristics, STI, motion monitoring, and tasks may be associated with an identification (ID) such as a UUID. Type 1, Type 2, or another device may acquire, store, retrieve, access, preprocess, condition, process, analyze, monitor, or apply TSCI. Type 1 and Type 2 devices may communicate network traffic on other channels (e.g., Ethernet, HDMI, USB, Bluetooth, BLE, WiFi, LTE, other networks, wireless multipath channels) in parallel with wireless signals. Type 2 devices may passively observe, monitor, or receive wireless signals from Type 1 devices on wireless multipath channels without establishing a connection (e.g., association / authentication) with or requesting service from the Type 1 device.

[0070] A transmitter (i.e., Type 1 device) can function (act as) a receiver (i.e., Type 2 device) temporarily, sporadically, continuously, repeatedly, interchangeably, alternately, simultaneously, in parallel, and / or simultaneously, and vice versa. A device may function as a Type 1 device (transmitter) and / or a Type 2 device (receiver) temporarily, sporadically, continuously, repeatedly, simultaneously, in parallel, and / or simultaneously. There may be multiple wireless nodes, each of which is a Type 1 (TX) and / or Type 2 (RX) device. TSCI may be obtained for each two nodes when exchanging / communicating wireless signals. An object's characteristics and / or STI may be monitored individually based on the TSCI or jointly based on two or more (e.g., all) TSCIs.

[0071] An object's movement may be monitored actively (in that its Type 1 device, Type 2 device, or both are wearable / associated with the object) and / or passively (in that both the Type 1 device and the Type 2 device are not wearable / associated with the object). It may be passive because the object may not be associated with a Type 1 device and / or a Type 2 device. The object (e.g., a user, an automated guided vehicle, or an AGV) may not need to carry / install any wearable / fixture (i.e., Type 1 devices and Type 2 devices are not wearable / attached equipment that the object needs to carry to perform a task). It may be active because the object may be associated with either a Type 1 device or a Type 2 device. The object may carry (or install) a wearable / attachment (e.g., a Type 1 device, a Type 2 device, or equipment communicatively coupled to either a Type 1 device or a Type 2 device).

[0072] The presentation may be visual, audio, image, video, animation, graphic presentation, text, etc. The computation of the task may be performed by a processor (or logic unit) of the Type 1 device, a processor (or logic unit) of the IC of the Type 1 device, a processor (or logic unit) of the Type 2 device, a processor of the IC of the Type 2 device, a local server, a cloud server, a data analysis subsystem, a signal analysis subsystem, and / or another processor. The task may be performed with or without a radio fingerprint or baseline (e.g., collection, processing, processing, transmission, and / or training phase / previous survey / latest survey / initial radio survey, passive indications), training, profile, trained profile, static profile, static profile, survey, initial radio survey, initial setup, installation, retraining, update, and reset).

[0073] A Type 1 device (TX device) may comprise at least one heterogeneous radio transmitter. A Type 2 device (RX device) may include at least one heterogeneous radio receiver. A Type 1 device and a Type 2 device may be co-located. A Type 1 device and a Type 2 device may be the same device. Any device may have a data processing unit / device, a computing unit / system, a network unit / system, a processor (e.g., a logic unit), a memory communicatively coupled to the processor, and a set of instructions stored in the memory to be executed by the processor. Some processors, memories, and sets of instructions may cooperate.

[0074] There may be multiple Type 1 devices interacting (e.g., communicating, exchanging signals / control / notifications / other data) with the same Type 2 device (or multiple Type 2 devices) and / or there may be multiple Type 2 devices interacting with the same Type 1 device. Multiple Type 1 / Type 2 devices may be synchronous and / or asynchronous, have the same / different window widths / sizes and / or time shifts, have the same / different synchronization start times, synchronization end times, etc. Wireless signals transmitted by multiple Type 1 devices may be sporadic, intermittent, continuous, repetitive, synchronous, simultaneous, and / or concurrent. Multiple Type 1 / Type 2 devices may operate independently and / or cooperatively. Type 1 and / or Type 2 devices may have / include heterogeneous hardware circuits (e.g., heterogeneous chips or ICs capable of generating / receiving wireless signals, extracting CI from received signals, or making CI available). They may be communicatively coupled to the same or different servers (e.g., cloud server, edge server, local server, hub device).

[0075] The operation of one device can be based on the operation, state, internal state, storage, processor, memory output, physical location, computational resources, or network of another device. Different devices may communicate directly and / or through another device / server / hub device / cloud server. A device can be associated with one or more users and have associated settings. Settings can be selected once, pre-programmed, and / or changed (e.g., adjusted, changed, modified) / varied over time. There may be additional steps in a method. Method steps and / or additional steps may be performed in the order shown or in a different order. Any steps may be performed in parallel, iteratively, or otherwise iteratively or otherwise. A user may be a human, adult, elderly adult, male, female, infant, child, baby, pet, animal, living being, machine, computer module / software, etc.

[0076] For one or more Type 1 devices interacting with one or more Type 2 devices, any processing (e.g., time domain, frequency domain) can be different for different devices. Processing can be based on location, orientation, direction, role, user-related characteristics, settings, configuration, available resources, available bandwidth, network connection, hardware, software, processor, co-processor, memory, battery life, available power, antenna, antenna type, antenna directional / unidirectional characteristics, power settings, and / or other parameters / characteristics of the device.

[0077] The wireless receiver (e.g., a Type 2 device) may receive a signal and / or another signal from the wireless transmitter (e.g., a Type 1 device). The wireless receiver may receive another signal from another wireless transmitter (e.g., a second Type 1 device). The wireless transmitter may transmit a signal and / or another signal to another wireless receiver (e.g., a second Type 2 device). The wireless transmitter, the wireless receiver, another wireless receiver, and / or another wireless transmitter may move with the object and / or another object. The other object may be tracked.

[0078] A Type 1 and / or Type 2 device may be capable of wirelessly coupling with at least two Type 2 and / or Type 1 devices. The Type 1 device may be triggered / controlled to switch / establish a wireless coupling (e.g., association, authentication) from the Type 2 device to a second Type 2 device at another location in the location. Similarly, the Type 2 device may be triggered / controlled to switch / establish a wireless coupling from the Type 1 device to a second Type 1 device at yet another location in the location. The switching may be controlled by a server (or hub device), a processor, the Type 1 device, the Type 2 device, and / or another device. The radios used before and after the switching may be different. A second wireless signal (second signal) may be transmitted through the channel between the Type 1 device and the second Type 2 device (or between the Type 2 device and the second Type 1 device). A second TSCI of the channel may be obtained from the second signal. The second signal may be the first signal. A characteristic, STI, and / or another quantity of the object may be monitored based on the second TSCI. The Type 1 and Type 2 devices may be the same. The characteristics, STIs, and / or other quantities with different timestamps may form a waveform. The waveform may be displayed in a presentation.

[0079] The wireless signal and / or another signal may have data embedded therein. The wireless signal may be a series of probe signals (e.g., repeated transmission of a probe signal, reuse of one or more probe signals). The probe signal may vary / change over time. The probe signal may be a standard-compliant signal, a protocol signal, a standardized wireless protocol signal, a control signal, a data signal, a wireless communication network signal, a cellular network signal, a WiFi signal, an LTE / 5G / 6G / 7G signal, a reference signal, a beacon signal, a motion detection signal, and / or a motion sensing signal. The probe signal may be formatted according to a wireless network standard (e.g., WiFi), a cellular network standard (e.g., LTE / 5G / 6G), or another standard. The probe signal may include a packet having a header and a payload. The probe signal may have data embedded therein. The payload may include data. The probe signal may replace a data signal. The probe signal may be embedded in a data signal. The wireless receiver, the wireless transmitter, the other wireless receiver and / or the other wireless transmitter may be associated with at least one processor, memory communicatively coupled to the respective processor, and / or a respective set of instructions stored in the memory that, when executed, cause the processor to perform any and / or all steps necessary to determine the object's STI (e.g., movement information), initial STI, initial time, direction, instantaneous position, instantaneous angle, and / or velocity.

[0080] The processor, memory, and / or set of instructions may be associated with a Type 1 device, at least one Type 2 device, an object, a device associated with the object, another device associated with the location, a cloud server, a hub device, and / or another server.

[0081] A Type 1 device can transmit a signal in a broadcast manner to at least one Type 2 device(s) over a location channel. The signal is transmitted without the Type 1 device establishing a wireless connection (e.g., association, authentication) with any Type 2 device and without the Type 2 device requesting service from the Type 1 device. A Type 1 device can transmit to a specific media access control (MAC) address common to multiple Type 2 devices. Each Type 2 device can tune its MAC address to a specific MAC address. The specific MAC address can be associated with a location. The association can be recorded in an association table in an association server (e.g., a hub device). A location can be identified by a Type 1 device, a Type 2 device, and / or another device based on the specific MAC address, a series of probe signals, and / or at least one TSCI extracted from the probe signals.

[0082] For example, a Type 2 device may be moved to a new location (e.g., from another location). A Type 1 device may be newly configured at a location such that the Type 1 and Type 2 devices are unaware of each other. During setup, the Type 1 device may be instructed / guided / triggered / controlled (e.g., using a dummy receiver, using a hardware pin configuration / connection, using a saved configuration, using a local configuration, using a remote configuration, using a downloaded configuration, using a hub device, or using a server) to send a series of probe signals to a specific MAC address. Upon powering up, the Type 2 device may scan for probe signals according to a table (e.g., stored in a designated source, server, hub device, cloud server) of MAC addresses that can be used to broadcast in different locations (e.g., different MAC addresses used for different locations such as a house, office, enclosure, floor, multi-story building, store, airport, mall, stadium, hall, station, subway, lot, region, area, district, province, city, country, continent, etc.). When the Type 2 device detects a probe signal sent to a specific MAC address, the Type 2 device may use the table to identify the location based on the MAC address.

[0083] The location of the Type 2 device may be calculated based on the specific MAC address, the series of probe signals, and / or at least one TSCI obtained by the Type 2 device from the probe signals. The calculation may be performed by the Type 2 device.

[0084] The specific MAC address may change (e.g., be adjusted, changed, modified) over time. It may change according to timetables, rules, policies, modes, conditions, circumstances, and / or changes. The specific MAC address may be selected based on MAC address availability, preselected lists, collision patterns, traffic patterns, data traffic between the Type 1 device and other devices, available bandwidth, random selection, and / or a MAC address switching plan. The specific MAC address may be the MAC address of a second wireless device (e.g., a dummy receiver or a receiver acting as a dummy receiver).

[0085] The Type 1 device may transmit a probe signal on a channel selected from the set of channels, and at least one CI of the selected channel may be obtained by each Type 2 device from the probe signal transmitted on the selected channel.

[0086] The selected channel may be changed (e.g., adjusted, changed, modified) over time. The change may be according to a timetable, rule, policy, mode, condition, situation, and / or change. The selected channel may be selected based on channel availability, random selection, a preselected list, co-channel interference, inter-channel interference, channel traffic patterns, data traffic between the Type 1 device and another device, effective bandwidth associated with the channel, security criteria, channel switching plans, criteria, quality criteria, signal quality conditions, and / or considerations.

[0087] The specific MAC address and / or selected channel information can be communicated between a Type 1 device and a server (e.g., a hub device) over a network. The specific MAC address and / or selected channel information can also be communicated between a Type 2 device and a server (e.g., a hub device) over another network. A Type 2 device can communicate the specific MAC address and / or selected channel information to another Type 2 device (e.g., via a mesh network, Bluetooth, WiFi, NFC, ZigBee, etc.). The specific MAC address and / or selected channel can be selected by a server (e.g., a hub device). The specific MAC address and / or selected channel can be signaled in an announcement channel by a Type 1 device, a Type 2 device, and / or a server (e.g., a hub device). Any information can be preprocessed before being communicated.

[0088] A wireless connection (e.g., association, authentication) between a Type 1 device and another wireless device can be established (e.g., using a signal handshake). The Type 1 device can send a first handshake signal (e.g., a sounding frame, a probe signal, a request to send RTS) to the other device. The other device can respond by sending a second handshake signal (e.g., a command or a clear to send CTS) to the Type 1 device, triggering the Type 1 device to send a signal (e.g., a series of probe signals) in a broadcast manner to multiple Type 2 devices without establishing a connection with any Type 2 devices. The second handshake signal can be a response or acknowledgment (e.g., an ACK) to the first handshake signal. The second handshake signal can include data having location and / or Type 1 device information. The other device can be a dummy device with a purpose (e.g., primary purpose, secondary purpose) to establish a wireless connection with the Type 1 device, receive the first signal, and / or send the second signal. The other device can be physically attached to the Type 1 device.

[0089] In another example, another device can send a third handshake signal to a Type-1 device that triggers the Type-1 device to broadcast a signal (e.g., a series of probe signals) to multiple Type-2 devices without establishing a connection (e.g., association, authentication) with any of the Type-2 devices. The Type-1 device can respond to the third special signal by sending a fourth handshake signal to the other device. Another device can be used to trigger multiple Type-1 devices to broadcast. The triggering may be sequential, partially sequential, partially parallel, or fully parallel. The other device may have multiple radio circuits to trigger multiple transmitters in parallel. Parallel triggering can also be achieved by using at least one additional device to perform a trigger in parallel with another device (as the other device does). The other device cannot communicate (or suspend communication) with the Type-1 device after establishing a connection with it. The suspended communication may resume. The other device can enter an inactive mode, dormant mode, sleep mode, standby mode, low power mode, OFF mode, and / or power-down mode after establishing a connection with the Type-1 device. The other device may have a specific MAC address such that the Type 1 device sends signals to the specific MAC address. The Type 1 device and / or the other device may be controlled and / or coordinated by a first processor associated with the Type 1 device, a second processor associated with the other device, a third processor associated with the specified source, and / or a fourth processor associated with the other device. The first and second processors may coordinate with each other.

[0090] A first series of probe signals may be transmitted by a first antenna of the Type 1 device to at least one first Type 2 device through a first channel at a first location. A second series of probe signals may be transmitted by a second antenna of the Type 1 device to at least one second Type 2 device through a second channel at a second location. The first series of probe signals and the second series of probe signals may be different / not different. The at least one first Type 2 device may be different / not different from the at least one second Type 2 device. The first and / or second series of probe signals may be broadcast without an established connection (e.g., association, authentication) between the Type 1 device and any Type 2 device. The first and second antennas may be the same / different.

[0091] The two locations may have different sizes, shapes, and multipath characteristics. The first and second locations may overlap. The immediate areas near the first and second antennas may overlap. The first and second channels may be the same or different. For example, the first may be WiFi and the second may be LTE. Or, both may be WiFi, but the first may be 2.4 GHz WiFi and the second may be 5 GHz WiFi. Or, both may be 2.4 GHz WiFi, but with different channel numbers, SSID names, and / or WiFi settings.

[0092] Each Type 2 device can obtain at least one TSCI from each series of probe signals, where a CI is each channel between the Type 2 device and the Type 1 device. Some first Type 2 device(s) and some second Type 2 device(s) may be the same. The first and second series of probe signals may be synchronous / asynchronous. The probe signals may be transmitted with data or may be replaced by data signals. The first antenna and the second antenna may be the same.

[0093] A first series of probe signals may be transmitted at a first rate (e.g., 30 Hz). A second series of probe signals may be transmitted at a second rate (e.g., 200 Hz). The first and second rates can be the same or different. The first and / or second rates can be changed (e.g., adjusted, varied, modified) over time. The change can be according to a timetable, rule, policy, mode, condition, situation, and / or change. Any rate can be changed (e.g., adjusted, varied, modified) over time.

[0094] The first and / or second series of probe signals may be transmitted to a first MAC address and / or a second MAC address, respectively. The two MAC addresses may be the same or different. The first series of probe signals may be transmitted in a first channel. The second series of probe signals may be transmitted in a second channel. The two channels may be the same or different. The first or second MAC address and the first or second channel may change over time. Any change may be according to a timetable, rule, policy, mode, state, condition, and / or change.

[0095] A Type 1 device and another device may control and / or coordinate, be physically attached to, or be of / within a common device. They may be controlled by / connected to a common data processor or connected to a common bus interconnect / network / LAN / Bluetooth network / NFC network / BLE / wired network / wireless network / mesh network / mobile network / cloud. They may share common memory or be associated with a common user, user device, profile, account, identity (ID), identifier, home, residence, physical address, location, geographic coordinates, IP subnet, SSID, home device, office device, and / or manufacturing device.

[0096] Each Type 1 device may be a signal source for a respective set of Type 2 devices (i.e., it sends a respective signal (e.g., a respective series of probe signals) to a respective set of Type 2 devices). Each Type 2 device selects a Type 1 device from all Type 1 devices as its signal source. Each Type 2 device may be selected asynchronously. At least one TSCI may be obtained by each Type 2 device from a respective series of probe signals from the Type 1 device, and the TSCI is a channel between the Type 2 device and the Type 1 device.

[0097] Each Type 2 device selects a Type 1 device as its signal source from among all Type 1 devices based on identity (ID) or Type 1 / Type 2 device identifier, task to be performed, past signal sources, history (e.g., of past signal sources, the Type 1 device, another Type 1 device, each Type 2 receiver, and / or another Type 2 receiver), threshold for switching signal sources, and / or user information, account, access information, parameters, characteristics, and / or signal strength (e.g., associated with the Type 1 device and / or each Type 2 receiver).

[0098] Initially, a Type 1 device may be a signal source for each of the initial set of Type 2 devices (i.e., the Type 1 device sends a respective signal (series of probe signals) to each of the initial set of Type 2 devices). Each initial respective Type 2 device selects a Type 1 device from among all Type 1 devices as its signal source.

[0099] A signal source of a particular Type 2 device (a Type 1 device) may be changed (e.g., adjusted, altered, modified) if: (1) the time interval between two adjacent probe signals (e.g., between the current probe signal and the immediately previous probe signal, or between the next probe signal and the current probe signal) received from the current signal source of the Type 2 device exceeds a first threshold; (2) the signal strength associated with the current signal source of the Type 2 device is less than a second threshold; (3) the processed signal strength associated with the current signal source of the Type 2 device is less than a third threshold, and the signal strength is processed with a low-pass filter, a band-pass filter, a median filter, a moving average filter, a weighted average filter, a linear filter, and / or a nonlinear filter; and / or (4) the signal strength (or processed signal strength) associated with the current signal source of the Type 2 device is below a fourth threshold for a significant percentage (e.g., 70%, 80%, 90%) of the recent time window. The percentage can exceed a fifth threshold. The first, second, third, fourth and / or fifth thresholds may be time-varying.

[0100] Condition (1) can occur when a Type 1 device and a Type 2 device gradually move farther away from each other, resulting in some probe signals from the Type 1 device becoming too weak to be received by the Type 2 device. Conditions (2) through (4) can occur when the two devices move far enough away from each other that the signal strength becomes very weak.

[0101] The signal source of a Type 2 device may not change if another Type 1 device has a signal strength weaker than the current signal source by a factor (eg, 1, 1.1, 1.2, or 1.5).

[0102] If the signal source is changed (adjusted, modified, modified, etc.), a new signal source may become effective in the near future (e.g., each next time). The new signal source may be the Type 1 device with the strongest signal strength and / or processed signal strength. The current signal source and the new signal source may be the same or different.

[0103] A list of available Type 1 devices may be initialized and maintained by each Type 2 device. The list may be updated by examining signal strengths and / or processed signal strengths associated with each set of Type 1 devices. A Type 2 device may select between a first series of probe signals from a first Type 1 device and a second series of probe signals from a second Type 1 device based on their respective probe signal rates, MAC addresses, channels, characteristics / properties / statuses, tasks to be performed by the Type 2 device, the first and second series of signal strengths, and / or other considerations.

[0104] The series of probe signals may be transmitted at a constant rate (e.g., 100 Hz). The series of probe signals may be scheduled at regular intervals (e.g., 0.01 seconds for 100 Hz), although each probe signal may experience small time perturbations, possibly due to timing requirements, timing control, network control, handshaking, message passing, collision avoidance, carrier sensing, congestion, resource availability, and / or other considerations.

[0105] The rate may be changed (e.g., adjusted, modified, or amended). The change may be according to a schedule (e.g., changed hourly), a rule, a policy, a mode, a condition, and / or a change (e.g., changed whenever an event occurs). For example, the rate may normally be 100 Hz, but may be changed to 1000 Hz in demanding situations, or to 1 Hz in low power / standby situations. The probe signal may be transmitted in bursts.

[0106] The probe signal rate may vary based on the task performed by the Type 1 or Type 2 device (e.g., a task may normally require 100 Hz, temporarily 1000 Hz for 20 seconds). In one example, transmitters (Type 1 devices), receivers (Type 2 devices), and associated tasks may be adaptively (and / or dynamically) associated with classes (e.g., classes that are low priority, high priority, emergency, critical, normal, privileged, non-subscribed, subscribed, paid, and / or unpaid). The (transmitter's) rate may be adjusted for some classes (e.g., high priority classes). If the needs of that class change, the rate can be changed (e.g., adjusted, modified, modified). If the receiver has critically low power, the rate may be reduced to reduce the receiver's power consumption for responding to the probe signal. In one example, the probe signal may be used to wirelessly transfer power to the receiver (Type 2 device), and the rate may be adjusted to control the amount of power transferred to the receiver.

[0107] The rate may be changed by (or based on): a server (e.g., a hub device), a Type 1 device, and / or a Type 2 device. Control signals may be communicated between them. The server may monitor, track, predict, and / or anticipate the needs of and / or tasks performed by the Type 2 device, and control the Type 1 device to change the rate. The server may make scheduled changes to the rate according to a timetable. The server may detect an emergency and immediately change the rate. The server may detect a developing condition and gradually adjust the rate.

[0108] Characteristics and / or STI (e.g., movement information) may be monitored individually based on TSCIs associated with a particular Type 1 device and a particular Type 2 device, and / or jointly based on any TSCIs associated with a particular Type 1 device and any Type 2 device, and / or jointly based on any TSCIs associated with a particular Type 2 device and any Type 1 device, and / or globally based on any TSCIs associated with any Type 1 device and any Type 2 device. Any joint monitoring may relate to: a user, a user account, a profile, a home, a map of the location, an environmental model of the location, and / or a user history, etc.

[0109] A first channel between a Type 1 device and a Type 2 device may be different from a second channel between another Type 1 device and another Type 2 device, and the two channels may be associated with different frequency bands, bandwidths, carrier frequencies, modulations, wireless standards, coding, encryption, payload characteristics, networks, network IDs, SSIDs, network characteristics, network settings, and / or network parameters, etc.

[0110] The two channels can relate to different types of wireless systems (e.g., two of the following: WiFi, LTE, LTE-A, LTE-U, 2.5G, 3G, 3.5G, 4G, Beyond 4G, 5G, 6G, 7G, cellular network standards, UMTS, 3GPP, GSM, EDGE, TDMA, FDMA, CDMA, WCDMA, TD-SCDMA, 802.11 systems, 802.15 systems, 802.16 systems, mesh networks, Zigbee, NFC, WiMax, Bluetooth, BLE, RFID, UWB, microwave systems, radar-like systems), for example, one channel being WiFi and the other being LTE.

[0111] The two channels may be associated with similar types of wireless systems, but different networks. For example, a first channel may be associated with a WiFi network named "Pizza and Pizza" in the 2.4 GHz band with a bandwidth of 20 MHz, while a second channel may be associated with a WiFi network with an SSID of "StarBud Hotspot" in the 5 GHz band with a bandwidth of 40 MHz. The two channels may be different channels within the same network (e.g., the "StarBud Hotspot" network).

[0112] In one embodiment, the wireless monitoring system can include training a classifier for multiple events at a location based on training TSCIs associated with the multiple events. The CIs or TSCIs associated with an event can take into account / include radio samples / characteristics / fingerprints associated with the event (and / or location, environment, object, object movement, state / emotional state / mental state / condition / stage / gesture / gait / action / movement / activity / daily activity / history / event of the object, etc.).

[0113] For each of a plurality of known events occurring at a location at a respective training (e.g., survey, wirelee survey, initial wireless survey) time associated with the known event, a respective training wireless signal (e.g., a respective series of training probe signals) may be transmitted to at least one first type 2 heterogeneous wireless device through a wireless multipath channel of the location at the respective training time by an antenna of the first type 1 heterogeneous wireless device using a processor, memory, and set of instructions of the first type 1 device.

[0114] At least one respective time series of training CIs (training TSCIs) may be asynchronously acquired by each of the at least one first Type 2 device from the (respective) training signal. The CIs may be CIs of a channel between the first Type 2 device and the first Type 1 device at a training time associated with a known event. The at least one training TSCI may be preprocessed. The training may be a wireless survey (e.g., during installation of the Type 1 device and / or the Type 2 device).

[0115] For a current event occurring within a location in a current period, a current wireless signal (e.g., a series of current probe signals) may be transmitted to at least one second Type 2 heterogeneous wireless device through a channel of the location in a current period related to the current event by an antenna of the second Type 1 heterogeneous wireless device using a processor, memory, and set of instructions of the second Type 1 device.

[0116] At least one time series of current CIs (current TSCIs) may be asynchronously acquired by each of the at least one second Type-2 device from a current signal (e.g., a series of current probe signals). The CIs may be CIs of a channel between the second Type-2 device and the second Type-1 device during a current time period associated with a current event. The at least one current TSCI may be preprocessed.

[0117] A classifier may be applied by at least one second Type 2 device to classify at least one current TSCI obtained from the series of current probe signals, to classify at least a portion of a particular current TSCI, and / or to classify a combination of at least a portion of a particular current TSCI with another portion of another TSCI. The classifier may divide the TSCIs (or features / STIs or other analysis or output responses) into clusters and associate the clusters with particular events / objects / subjects / locations / movements / activities. Labels / tags may be generated for the clusters. The clusters may be stored and searched. A classifier may be applied to associate the current TSCI (or perhaps a characteristic / STI or other analysis / output response related to the current event) with: a cluster, a known / specific event, a class / category / group / grouping / list / cluster, a set of known events / subject / location / movement / activity, an unknown event, a class / category / group / grouping / list / cluster, a set of unknown events / subject / location / movement / activity, and / or another event / subject / location / movement / activity / class / category / group / grouping / list / cluster. Each TSCI may include at least one CI, each associated with a respective timestamp. Two TSCIs associated with two Type 2 devices differ by different start times, durations, stop times, amount of CIs, sampling frequencies, and sampling periods. The CIs may have different characteristics. The first and second Type 1 devices may be at the same location. They may be the same device. At least one second Type 2 device (or their location) may be a replacement for at least one first Type 2 device (or their location). The particular second Type 2 device and the particular first Type 2 device may be the same device.

[0118] The subset of first Type 2 devices and the subset of second Type 2 devices may be the same. At least one second Type 2 device and / or at least one subset of second Type 2 devices may be a subset of at least one first Type 2 device. At least one first Type 2 device and / or at least one subset of first Type 2 devices may be a replacement for a subset of at least one second Type 2 device. At least one second Type 2 device and / or at least one subset of second Type 2 devices may be a replacement for a subset of at least one first Type 2 device. At least one second Type 2 device and / or at least one subset of second Type 2 devices may be in the same respective locations as the subset of at least one first Type 2 device. At least one first Type 2 device and / or at least one subset of first Type 2 devices may be in the same respective locations as the subset of at least one second Type 2 device.

[0119] The antenna of the Type 1 device and the antenna of the second Type 1 device may be in the same location. The antenna of the at least one second Type 2 device and / or the antenna of the subset of at least one second Type 2 device may be in the same respective location as each antenna of the subset of at least one first Type 2 device. The antenna of the at least one first Type 2 device and / or the antenna of the subset of at least one first Type 2 device may be in the same respective location as each antenna of the subset of at least one second Type 2 device.

[0120] A first section of a first duration of a first TSCI and a second section of a second duration of a second TSCI may be aligned. A map between items in the first section and items in the second section may be calculated. The first section may include a first segment (e.g., a subset) of the first TSCI having a first start / end time and / or another segment (e.g., a subset) of the processed first TSCI. The processed first TSCI may be the first TSCI processed by the first operation. The second section may include a second segment (e.g., a subset) of the second TSCI having a second start time and a second end time, and another segment (e.g., a subset) of the processed second TSCI. The processed second TSCI may be the second TSCI processed by the second operation. The first operation and / or the second operation may include subsampling, resampling, interpolation, filtering, transformation, feature extraction, preprocessing, and / or other operations.

[0121] A first item in a first section may be mapped to a second item in a second section. A first item in a first section may also be mapped to another item in a second section. Another item in a first section may also be mapped to a second item in a second section. The mapping may be one-to-one, one-to-many, many-to-one, or many-to-many. At least one function of at least one of the first item in a first section of a first TSCI, another item in a first TSCI, a timestamp of the first item, a time difference of the first item, a time difference of the first item, an adjacent timestamp of the first item, an adjacent timestamp of the first item, another timestamp related to the first item, a second item in a second section of a second TSCI, another item in a second TSCI, a timestamp of the second item, a time difference of the second item, a time difference of the second item, an adjacent timestamp of the second item, and another timestamp related to the second item may satisfy at least one constraint.

[0122] One constraint may be that the difference between the timestamp of the first item and the timestamp of the second item may be bounded upper by an adaptive (and / or dynamically adjusted) upper threshold and lower by an adaptive lower threshold.

[0123] The first section may be the entire first TSCI. The second section may be the entire second TSCI. The first time duration may be equal to the second time duration. The time duration sections of a TSCI may be determined adaptively (and / or dynamically). A provisional section of the TSCI may be calculated. A start time and an end time of a section (e.g., provisional section, section) may be determined. A section may be determined by removing the start and end portions of the provisional section. The start portion of the provisional section may be determined as follows: Iteratively, items in the provisional section with increasing timestamps may be considered as the current item, one item at a time.

[0124] At each iteration, at least one activity measure / index may be calculated and / or considered. The at least one activity measure may be associated with at least one of the following: a current item associated with the current timestamp, a past item in the provisional section with a timestamp not greater than the current timestamp, and / or a future item in the provisional section with a timestamp not less than the current timestamp. If at least one criterion (e.g., quality criterion, signal quality condition) associated with the at least one activity measure is met, the current item may be added to the beginning of the provisional section.

[0125] The at least one criterion associated with the activity measure may include at least one of the following: (a) the activity measure is less than an adaptive (dynamically adjusted) upper threshold, (b) the activity measure is greater than an adaptive lower threshold, (c) the activity measure is less than an adaptive upper threshold consecutively for at least a predetermined amount of consecutive timestamps, (d) the activity measure is greater than an adaptive lower threshold consecutively for at least another predetermined amount of consecutive timestamps, (e) the activity measure is less than an adaptive upper threshold consecutively for at least a predetermined percentage of a predetermined amount of consecutive timestamps, (f) the activity measure is greater than an adaptive lower threshold consecutively for at least another predetermined percentage of another predetermined amount of consecutive timestamps, (g) another activity measure associated with another timestamp related to the current timestamp is less than another adaptive upper threshold and greater than another adaptive lower threshold, or (h) at least one activity measure associated with at least one respective timestamp related to the current timestamp is less than a respective upper threshold and greater than a respective lower threshold. (i) the percentage of timestamps associated with the activity measures in a set of timestamps associated with the current timestamp that are less than their respective upper thresholds and greater than their respective lower thresholds exceeds a threshold; and (j) another criterion (e.g., quality criterion, signal quality condition).

[0126] The activity measure / index associated with the item at time T1 may include at least one of the following: (1) a first function of the item at time T1 and the item at time T1-D1, where D1 is a predetermined positive quantity (e.g., a fixed time offset); (2) a second function of the item at time T1 and the item at time T1+D1; (3) a third function of the item at time T1 and the item at time T2, where T2 is a predetermined quantity (e.g., a fixed initial reference time; T2 may change (e.g., adjusted, varied, modified) over time; T2 may be updated periodically; T2 may be the start of a period and T1 may be a sliding time in the period); and (4) a fourth function of the item at time T1 and other items.

[0127] At least one of the first function, the second function, the third function, and / or the fourth function may be a function (e.g., F(X, Y, ...)) with at least two arguments X and Y. The two arguments may be scalars. The function (e.g., F) may be at least one of X, Y, (XY), (YX), abs(XY), X^a, Y^b, abs(X^aY^b), (XY)^a, (X / Y), (X+a) / (Y+b), (X^a / Y^b), and ((X / Y)^ab), where a and b may be certain predetermined quantities. For example, the function may be simply abs(XY), or (XY)^2, (XY)^4. The function may be a robust function. For example, the function is (XY)^2 when abs(XY) is less than a threshold T, and (XY)+a when abs(XY) is greater than T. Alternatively, the function may be a constant when abs(XY) is greater than T. Also, when abs(Xy) is greater than T, the function may be bounded by a slowly increasing function so that outliers cannot significantly affect the results. Another example of this function may be (abs(X / Y)-a) (where a=1). In this way, when X=Y (i.e., no change or activity), the function yields a value of 0. When X is greater than Y, (X / Y) is greater than 1 (assuming X and Y are positive), and the function is positive. When X is less than Y, (X / Y) is less than 1, and the function is negative. In another example, both arguments X and Y may be n-tuples, with X = (x_1, x_2, ..., x_n) and Y = (y_1, y_2, ..., y_n). The function may be at least one of x_i, y_i, (x_i - y_i), (y_ix_i), abs(x_i - y_i), x_i^a, y_i^b, abs(x_i^a - y_i^b), (x_i - y_i)^a, (x_i / y_i), (x_i + a) / (y_i + b), (x_i^a / y_i^b), and ((x_i / y_i)^ab), where i is a component index of the n-tuples X and Y, and 1 <= i <= n.For example, the component index of x_1 is i = 1, the component index of x_2 is i = 2, etc. The function can include a sum of another function per component of at least one of x_i, y_i, (x_i-y_i), (y_ix_i), abs(x_i-y_i), x_i^a, y_i^b, abs(x_i^a-y_i^b), (x_i-y_i)^a, (x_i / y_i), (x_i+a) / (y_i +b), (x_i^a / y_i^b), and ((x_i / y_i)^ab), where i is a component index of the n-tuples X and Y. For example, this function could be of the form sum_{i=1}^n(abs(x_i / y_i)-1) / n, or sum_{i=1}^nw_i*(abs(x_i / y_i)-1), where w_i is the weight of component i.

[0128] The map may be computed using dynamic time warping (DTW). The DTW may include constraints on at least one of the map, the items of the first TSCI, the items of the second TSCI, the first duration, the second duration, the first section, and / or the second section. Suppose the i^{th} domain item is mapped to the j^{th} range item in the map. The constraints may be on the allowable combinations of i and j (constraints on the relationship between i and j). A mismatch cost between the first section of the first duration of the first TSCI and the second section of the second duration of the second TSCI may be computed.

[0129] The first section and the second section may be aligned such that a map including a plurality of links may be established between a first item of the first TSCI and a second item of the second TSCI. Each link may associate one of the first items with a first timestamp and one of the second items with a second timestamp. A mismatch cost between the aligned first section and the aligned second section may be calculated. The mismatch cost may include a function of a cost for an item between the first item and the second item associated by a particular link of the map and a link-wise cost for a link associated with the particular link of the map.

[0130] The aligned first section and the aligned second section may be represented as a first vector and a second vector, respectively, of the same vector length. The mismatch cost may include at least one of a dot product, a dot product-like measure, a correlation-based measure, a correlation indicator, a covariance-based measure, a discrimination score, a distance, a Euclidean distance, an absolute distance, an Lk distance (e.g., L1, L2,...), a weighted distance, a distance-like measure, and / or another similarity value between the first vector and the second vector. The mismatch cost may be normalized by the respective vector lengths.

[0131] A parameter derived from the mismatch cost between a first section of a first duration of a first TSCI and a second section of a second duration of a second TSCI can be modeled with a statistical distribution, and at least one of a scale parameter, a location parameter, and / or another parameter of the statistical distribution can be estimated.

[0132] The first section of the first duration of the first TSCI may be a sliding section of the first TSCI, and the second section of the second duration of the second TSCI may be a sliding section of the second TSCI.

[0133] A first sliding window may be applied to a first TSCI, and a corresponding second sliding window may be applied to a second TSCI, and the first sliding window of the first TSCI and the corresponding second sliding window of the second TSCI may be aligned.

[0134] A mismatch cost between the aligned first sliding window of the first TSCI and the corresponding aligned second sliding window of the second TSCI can be calculated, and the current event may be associated with at least one of a known event, an unknown event, and / or another event based on the mismatch cost.

[0135] The classifier may be applied to at least one of each first section of a first duration of the first TSCI and / or each second section of a second duration of the second TSCI to obtain at least one provisional classification result, each provisional classification result being associated with a respective first section and a respective second section.

[0136] The current event can be associated with at least one of a known event, an unknown event, a class / category / group / grouping / list / set of unknown events, and / or another event based on the mismatch cost. The current event can be associated with at least one of a known event, an unknown event, and / or another event based on the most numerous provisional classification results in the multiple sections of the first TSCI and the multiple sections of the corresponding second TSCI. For example, if the mismatch cost points to a specific known event for N consecutive times (e.g., N=10), the current event can be associated with the specific known event. In another example, the current event can be associated with the specific known event if the percentage of mismatch costs within the immediately preceding N consecutive times that point to the specific known event exceeds a predetermined threshold (e.g., >80%).

[0137] In another example, the current event may be associated with a known event that achieves the smallest mismatch cost the most times in time. The current event may be associated with a known event that achieves the smallest overall mismatch cost, which is a weighted average of at least one mismatch cost associated with at least one first section. The current event may be associated with a particular known event that achieves another minimum overall cost. The current event may be associated with an "unknown event" if none of the known events achieves a mismatch cost below a first threshold T1 in a sufficient percentage of at least one first section. The current event may also be associated with an "unknown event" if none of the known events achieves an overall mismatch cost below a second threshold T2. The current event may be associated with at least one of a known event, an unknown event, and / or another event based on the mismatch cost and additional mismatch cost associated with at least one additional section of the first TSCI and at least one additional section of the second TSCI. The known events may include at least one of a door closed event, a door open event, a window closed event, a window open event, a multi-state event, an on state event, an off state event, an intermediate state event, a continuous state event, a discrete state event, a person present event, a person absent event, a life present event, and / or a life absent event.

[0138] A projection for each CI may be trained using a dimensionality reduction method based on the training TSCI. The dimensionality reduction method may include at least one of principal component analysis (PCA), PCA with different kernels, independent component analysis (ICA), Fisher's linear discriminant, vector quantization, supervised learning, unsupervised learning, self-organizing maps, autoencoders, neural networks, deep neural networks, and / or another method. The projection may be applied to at least one of the training TSCI associated with at least one event and / or the current TSCI for the classifier.

[0139] A classifier for at least one event may be trained based on a projection associated with the at least one event and associated training TSCI. At least one current TSCI may be classified / categorized based on the projection and the current TSCI. The projection may be retrained using at least one of a dimensionality reduction method and another dimensionality reduction method based on at least one of the training TSCI, the at least one current TSCI before retraining the projection, and / or additional training TSCI. The other dimensionality reduction methods may include at least one of principal component analysis (PCA), PCA with different kernels, independent component analysis (ICA), Fisher's linear discriminant, vector quantization, supervised learning, unsupervised learning, self-organizing maps, autoencoders, neural networks, deep neural networks, and / or yet another method. A classifier for at least one event may be retrained based on at least one of the retrained projection, the training TSCI associated with the at least one event, and / or the at least one current TSCI. At least one current TSCI may be classified based on the retrained projections, the retrained classifier, and / or the current TSCI.

[0140] Each CI may include a vector of complex values. Each complex value may be preprocessed to provide a magnitude of the complex value. Each CI may be preprocessed to provide a vector of non-negative real numbers containing the magnitude of the corresponding complex value. Each training TSCI may be weighted in training the projection. The projection may include multiple projection components. The projection may include at least one most significant projection component. The projection may include at least one projected component that may be useful to the classifier.

[0141] Channel / Channel information / Venue / Spatia-temporal information / Motion / Object

[0142] Channel information (CI) may include signal strength, signal amplitude, signal phase, spectral power measures, modem parameters (e.g., used in connection with modulation / demodulation in digital communication systems such as WiFi, 4G / LTE, etc.), dynamic beamforming information (including feedback or steering matrices generated by wireless communication devices according to standardization processes such as IEEE 802.11 or other standards), transfer function components, radio conditions (e.g., used in digital communication systems to decode digital data, baseband processing conditions, RF processing conditions, etc.), measurable variables, sensing data, coarse-grained / fine-grained information for layers (e.g., physical layer, data link layer, MAC layer, etc.), digital settings, gain settings, RF filter settings, RF front-end switch settings, DC offset settings, DC correction settings, IQ correction settings, effects of the environment (e.g., location) on the radio signal during propagation, input signal (radio signal transmitted by a Type 1 device) and output signal (radio signal transmitted by a Type 2 device) The CI may be associated with or include a transformation of a CI into a wireless signal (e.g., wireless signal received by the wireless communication device), steady state behavior of the environment, a condition profile, wireless channel measurements, a received signal strength indicator (RSSI), channel state information (CSI), beamforming dynamic information (including feedback or steering matrices generated by the wireless communication device according to a standardization process such as IEEE 802.11 or other standards), a channel impulse response (CIR), a channel frequency response (CFR), characteristics of frequency components (e.g., subcarriers) in a bandwidth, channel characteristics, a channel response, a timestamp, aiding information, data, metadata, user data, account data, access data, security data, session data, status data, supervisory data, home data, identification (ID), identifier, device data, network data, proximity data, environmental data, real-time data, sensor data, stored data, encrypted data, compressed data, protected data, and / or other channel information. Each CI may be associated with a timestamp and / or a time of arrival.CSI is used to equalize / restore / minimize / reduce multipath channel effects (transmission channels) and demodulate signals similar to those transmitted by a transmitter through the multipath channel. CI can be associated with information related to the frequency band, frequency signature, frequency phase, frequency amplitude, frequency trend, frequency characteristics, frequency-like characteristics, time-domain elements, frequency-domain elements, time-frequency domain elements, orthogonal decomposition characteristics, and / or non-orthogonal decomposition characteristics of the signal passing through the channel. TSCI can be a stream of wireless signals (e.g., CI).

[0143] The CI may be pre-processed, processed, post-processed, stored (e.g., in a local memory, portable / mobile memory, removable memory, storage network, cloud memory, in a volatile manner, in a non-volatile manner), retrieved, transmitted, and / or received. One or more modem parameters and / or radio condition parameters may be kept constant. The modem parameters may be applied to a radio subsystem. The modem parameters may represent radio conditions. A motion detection signal (e.g., a baseband signal and / or packets decoded / demodulated from the baseband signal, etc.) may be obtained by processing (e.g., downconverting) a first radio signal (e.g., an RF / WiFi / LTE / 5G signal) by the radio subsystem using the radio conditions represented by the stored modem parameters. The modem parameters / radio conditions may be updated (e.g., using previous modem parameters or previous radio conditions). Both the previous and updated modem parameters / radio conditions may be applied to a radio subsystem of a digital communication system. Both the previous and updated modem parameters / radio conditions may be compared / analyzed / processed / monitored in a task.

[0144] The channel information may also be modem parameters (e.g., stored or newly calculated) used to process the wireless signal. The wireless signal may include multiple probe signals. The same modem parameters can be used to process multiple probe signals. The same modem parameters can also be used to process multiple wireless signals. The modem parameters may include parameters indicating settings or overall configurations for operation of the radio subsystem or baseband subsystem (or both) of the wireless sensor device. The modem parameters may include one or more of gain settings, RF filter settings, RF front-end switch settings, DC offset settings, or IQ compensation settings for the radio subsystem, or digital DC correction settings, digital gain settings, and / or digital filtering settings (e.g., for the baseband subsystem). CI may also relate to information related to time, time signature, timestamp, time amplitude, time phase, time trend, and / or time characteristics of a signal. CI may be associated with information related to the time-frequency division, signature, amplitude, phase, trend, and / or characteristics of a signal. CI may relate to signal decomposition. A CI may relate to information related to direction, angle of arrival (AoA), angle of a directional antenna, and / or phase of a signal passing through a channel. A CI may relate to the attenuation pattern of a signal passing through a channel. Each CI may be associated with a Type 1 device and a Type 2 device. Each CI may be associated with an antenna of a Type 1 device and an antenna of a Type 2 device.

[0145] The CI can be obtained from communication hardware (e.g., a Type 2 device or a Type 1 device) capable of providing the CI. The communication hardware can be a WiFi-enabled chip / IC (integrated circuit), a chip compliant with 802.11 or 802.16 or other wireless / wireless standards, a next-generation WiFi-enabled chip, an LTE-enabled chip, a 5G-enabled chip, a 6G / 7G / 8G-enabled chip, a Bluetooth-enabled chip, an NFC (near field communication)-enabled chip, a BLE (Bluetooth low energy)-enabled chip, a UWB chip, or other communication chips (e.g., Zigbee, WiMax, mesh networks). The communication hardware calculates the CI and stores the CI in a buffer memory so that the CI is available for extraction. The CI can include data related to channel state information (CSI) and / or at least one matrix. The at least one matrix can be used for channel equalization, beamforming, etc. The channel can be associated with a location. Attenuation can be due to signal propagation at the location, signal propagation through / at / near the air (e.g., the air at the location), reflection, refraction, diffraction, refractive media / reflective surfaces such as walls, doors, furniture, obstacles, and / or barriers, etc. Attenuation can be due to reflections off surfaces and obstacles (e.g., reflective surfaces, obstacles) such as floors, ceilings, furniture, fixtures, objects, people, pets, etc. Each CI can be associated with a timestamp. Each CI can include N components (e.g., N frequency-domain components in CFR, N time-domain components in CIR, or N decomposed components). Each component can be associated with a component index. Each component can be a real, imaginary, or complex quantity, magnitude, phase, flag, and / or set. Each CI can include a vector or matrix of complex numbers, a set of mixed quantities, and / or a multidimensional collection of at least one complex number.

[0146] Components of the TSCI associated with a particular component index may form respective component time series associated with the respective index. The TSCI may be divided into N component time series. Each individual component time series is associated with a respective component index. Object motion characteristics / STI may be monitored based on the component time series. In one example, one or more ranges of CI components (e.g., one range from component 11 to component 23, a second range from component 44 to component 50, and a third range having only one component) may be selected based on a criterion / cost function / signal quality metric (e.g., based on signal-to-noise ratio and / or interference level) for further processing.

[0147] Component-wise characteristics of the TSCI component-feature time series may be calculated. The component-wise characteristics may be scalars (e.g., energy) or functions with domains and ranges (e.g., autocorrelation functions, transforms, inverse transforms). Object motion characteristics / STIs may be monitored based on the component-wise characteristics. Overall characteristics (e.g., aggregate characteristics) of the TSCI may be calculated based on the component-wise characteristics of each TSCI component time series. The overall characteristics may be a weighted average of the component-wise characteristics. Object motion characteristics / STIs may be monitored based on the overall characteristics. A total quantity may be a weighted average of the individual quantities.

[0148] Type 1 devices and Type 2 devices may support WiFi, WiMax, 3G / 3G Beyond, 4G / 4G Beyond, LTE, LTE-A, 5G, 6G, 7G, Bluetooth, NFC, BLE, Zigbee, UWB, UMTS, 3GPP, GSM, EDGE, TDMA, FDMA, CDMA, WCDMA, TD-SCDMA, mesh networks, proprietary wireless systems, IEEE 802.11 standards, 802.15 standards, 802.16 standards, 3GPP standards, and / or other wireless systems.

[0149] A common wireless system and / or a common wireless channel may be shared by a Type 1 transceiver and / or at least one Type 2 transceiver. The at least one Type 2 transceiver may transmit respective signals simultaneously (or: asynchronously, synchronously, sporadically, continuously, repeatedly, in parallel, simultaneously, and / or at one time) using the common wireless system and / or the common wireless channel. The Type 1 transceiver may transmit signals to the at least one Type 2 transceiver using the common wireless system and / or the common wireless channel.

[0150] Each Type 1 device and Type 2 device may have at least one transmit and receive antenna. Each CI may be associated with one of the transmit antennas of the Type 1 device and one of the receive antennas of the Type 2 device. Each pair of transmit and receive antennas may be associated with a link, path, communication path, signal hardware path, etc. For example, if the Type 1 device has M (e.g., 3) transmit antennas and the Type 2 device has N (e.g., 2) receive antennas, there may be M×N (e.g., 3×2=6) links or paths. Each link or path may be associated with a TSCI.

[0151] At least one TSCI may correspond to various antenna pairs between a Type 1 device and a Type 2 device. The Type 1 device may have at least one antenna. The Type 2 device may also have at least one antenna. Each TSCI may be associated with an antenna of the Type 1 device and an antenna of the Type 2 device. Averaging or weighted averaging across antenna links may be performed. The averaging or weighted averaging may be across at least one TSCI. The averaging may optionally be performed over a subset of the at least one TSCI corresponding to a subset of the antenna pairs.

[0152] The timestamps of some CIs of a TSCI may be irregular and may be corrected so that the corrected timestamps of the time-corrected CIs are evenly spaced in time. In the case of multiple Type 1 devices and / or multiple Type 2 devices, the corrected timestamps may be related to the same clock or different clocks. An original timestamp associated with each of the CIs may be determined. The original timestamps may not be evenly spaced in time. The original timestamps of all CIs of a particular portion of a particular TSCI in the current sliding time window may be corrected so that the corrected timestamps of the time-corrected CIs are evenly spaced in time.

[0153] Characteristics and / or STI (e.g., motion information) may include location, location coordinates, change in location, location (e.g., initial location, new location), location on map, height, horizontal position, vertical position, distance, displacement, speed, acceleration, rotational speed, rotational acceleration, direction, movement angle, orientation, direction of movement, rotation, path, deformation, translation, contraction, extension, gait, gait cycle, head movement, repetitive movement, periodic movement, pseudo-periodic movement, impulse movement, sudden movement, falling movement, transient movement, behavior, transient behavior, movement cycle, movement frequency, time trend, temporal profile, temporal characteristics, occurrence, change, temporal change, change in CI, change in frequency, change in timing, change in gait cycle, timing, start time, start time, end time, duration, movement history, movement type, movement classification, frequency, frequency spectrum, frequency characteristics, presence, absence, proximity, proximity , retreat, object identification / identifier, object composition, head movement velocity, head movement direction, mouth-related rate, eye-related rate, breathing rate, heart rate, tidal volume, breathing depth, inhalation time, exhalation time, inhalation to sweep time ratio, airflow rate, heart rate interval, heart rate variability, hand movement rate, hand movement direction, leg movement, body movement, walking speed, hand movement velocity, position characteristics, object movement-related characteristics (e.g., change in position / location), tool movement, machine movement, compound movement, and / or combination of multiple movements, event, signal statistics, signal dynamics, anomaly, movement statistics, movement parameters, motion detection indication, motion magnitude, motion phase, similarity score, distance score, Euclidean distance, weighted distance, L_1 norm, L_2 norm, L_k norm for k>2, statistical distance, correlation, correlation indicator,Autocorrelation, covariance, autocovariance, cross-covariance, inner product, Cartesian product, motion signal transformation, motion features, motion presence, motion absence, motion localization, motion discrimination, motion recognition, object presence, object absence, object entrance, object exit, object change, movement cycle, movement count, gait cycle, movement rhythm, movement deformation, gesture, handwriting, head movement, mouth movement, cardiac movement, visceral movement, motion trend, size, length, area, volume, volume, shape, morphology, tag, start / start position, end position, start / start amount, end amount, event, fall event, security The information may include security events, accident events, home events, office events, factory events, warehouse events, manufacturing events, assembly line events, maintenance events, car-related events, navigation events, tracking events, door events, door open events, door close events, window events, window open events, window closed events, repeatable events, one-time events, consumption, non-consumption, status, physical status, health status, comfort status, emotional status, mental status, other events, analysis, output response, and / or other information. Characteristics and / or STIs may be calculated / monitored based on features calculated from the CI or TSCI (e.g., feature calculation / extraction). Static segments or profiles (and / or dynamic segments / profiles) may be identified / calculated / analyzed / monitored / extracted / acquired / marked / presented / indicated / highlighted / stored / communicated based on feature analysis. Analysis may include motion detection / motion assessment / presence detection. Computational workloads may be shared among Type 1 devices, Type 2 devices, and other processors.

[0154] The Type 1 device and / or the Type 2 device may be a local device, which may be a smartphone, a smart device, a TV, a sound bar, a set-top box, an access point, a router, a repeater, a wireless signal repeater / extender, a remote control, a speaker, a fan, a refrigerator, a microwave oven, a coffee machine, a hot water pot, an appliance, a table, a chair, a light, a lamp, a door lock, a camera, a microphone, a motion sensor, a security device, a fire hydrant, a garage door switch, a power adapter, a computer, a dongle, a computer peripheral, an electronic pad, a sofa, a tile, an accessory, a home device, a vehicle device, an office device, a building equipment, a manufacturing equipment, a watch, a glass, a clock, a television, an oven, an air conditioner, an accessory, a utility, an appliance, a smart machine, a smart vehicle, an Internet of Things (IoT), a smart house, a smart office, a smart building, a smart parking lot, a smart system, and other devices.

[0155] Each Type 1 device may be associated with a respective identifier (e.g., ID). Each Type 2 device may also be associated with a respective identification (ID). The ID may include numbers, a combination of text and numbers, a name, a password, an account, an account ID, a web link, a web address, an index to some information, and / or another ID. The ID may be assigned. The ID may be assigned by hardware (e.g., hardwired, via a dongle, and / or other hardware), software, and / or firmware. The ID may be stored (e.g., in a database, in memory, in a server (e.g., a hub device), in the cloud, locally stored, remotely stored, permanently stored, or temporarily stored) and may be searched. The ID may be associated with at least one record, account, user, household, address, phone number, social security number, customer number, another ID, another identifier, timestamp, and / or collection of data. The ID and / or a portion of the ID of the Type 1 device may be made available to the Type 2 device. The ID may be used by Type 1 devices and / or Type 2 devices for registration, initialization, communication, identification, verification, detection, recognition, authentication, access control, cloud access, networking, social networking, logging, recording, cataloging, classification, tagging, association, pairing, transactions, electronic transactions, and / or intellectual property control.

[0156] The object may be a person, user, subject, passenger, child, elderly, infant, sleeping infant, infant in a vehicle, patient, worker, high value worker, expert, medical specialist, waiter, customer in a mall, traveller at an airport / train station / bus terminal / shipping terminal, staff / laborer / customer service person in a factory / mall / supermarket / office / workplace, service person in a sewer / air ventilation system / lift well, lift in a lift well, elevator, inmate, person to be tracked / monitored, animal, plant, living thing, pet, dog, cat, smartphone, phone accessory, computer, tablet, portable computer, dongle, computer accessory equipment, network equipment, WiFi equipment, IoT device, smart watch, smart glasses, smart device, speaker, key, smart key, wallet, wallet, handbag, backpack, goods, cargo, luggage, equipment, motor, machine, air conditioner, fan, air conditioning equipment, lighting fixture, movable light, television, camera, audio and / or video equipment, It can be stationery, surveillance equipment, parts, signs, tools, carts, tickets, parking passes, passes, plane tickets, credit cards, plastic cards, access cards, food packaging, utensils, tables, chairs, cleaning equipment / tools, vehicles, automobiles, cars in parking facilities, goods in a warehouse / store / supermarket / distribution center, boats, bicycles, airplanes, drones, remote controlled cars / planes / boats, robots, manufacturing equipment, assembly lines, materials / unfinished parts / robots / trolleys / transports on a factory floor, tracked objects in an airport / shopping mart / supermarket, non-objects, absence of objects, presence of objects, objects with shape, objects changing shape, shapeless objects, mass of a fluid, mass of a liquid, mass of a gas / smoke, fire, flames, electromagnetic (EM) sources, EM media, and / or other objects.

[0157] The object itself may be communicatively coupled to several networks, such as WiFi, MiFi, 3G / 4G / LTE / 5G / 6G / 7G, Bluetooth, NFC, BLE, WiMax, Zigbee, UMTS, 3GPP, GSM, EDGE, TDMA, FDMA, CDMA, WCDMA, TD-SCDMA, mesh networks, ad-hoc networks, and / or other networks. The object itself may be AC-powered and bulky, and may be moved during installation, cleaning, maintenance, renovation, etc. The object may also be placed on a mobile platform, such as a lift, pad, mobile platform, elevator, conveyor belt, robot, drone, forklift, car, boat, or vehicle. The object may have multiple parts, each with a different motion (e.g., change of location / position). For example, the object may be a person walking in front. While walking, his left and right hands may move in different directions with different instantaneous speeds, accelerations, and motions.

[0158] The wireless transmitter (e.g., a Type 1 device), the wireless receiver (e.g., a Type 2 device), another wireless transmitter, and / or another wireless receiver may travel with the object and / or another object (e.g., in a previous trip, a current trip, and / or a future trip). They may be communicatively coupled to one or more nearby devices. They may transmit TSCIs and / or information related to TSCIs to nearby devices and / or to each other. They may be associated with nearby devices. The wireless transmitter and / or wireless receiver may be part of a small (e.g., coin-sized, cigarette-pack-sized, or even smaller) lightweight portable device. The portable device may be wirelessly coupled to the nearby device.

[0159] The nearby device may be a smartphone, an iPhone, an Android phone, a smart device, a smart appliance, a smart vehicle, a smart gadget, a smart TV, a smart refrigerator, a smart speaker, a smart watch, smart glasses, a smart pad, an iPad, a computer, a wearable computer, a notebook computer, a gateway. The nearby device may be connected to a cloud server, a local server (e.g., a hub device), and / or other servers via the Internet, a wired Internet connection, and / or a wireless Internet connection. The nearby device may be portable. The portable device, nearby devices, local server (e.g., hub device), and / or cloud server can share tasks (e.g., acquiring TSCI, determining object characteristics / STI related to object movement (e.g., position / change in position), calculating time series of power (e.g., signal strength) information, determining / calculating specific functions, searching for local extrema, classification, identifying specific values ​​of offset time, denoising, processing, simplification, cleaning, wireless smart sensing tasks, extracting CIs from signals, switching, segmenting, estimating trajectory / path / track, processing maps, processing trajectory / path / track based on environmental models / constraints / limits, correction, correction adjustment, tuning, map-based (or model-based) correction, error detection, checking for boundary hits, thresholding) and calculation and / or storage of information (e.g., TSCI). Nearby devices may move / not move with the object. Nearby devices can be portable / non-portable / mobile / non-mobile. Nearby devices can use battery power, solar, AC power, and / or other power sources. The nearby device may have replaceable / non-replaceable and / or rechargeable / non-rechargeable batteries. The nearby device may resemble the object. The nearby device may have the same (and / or similar) hardware and / or software as the object.The nearby devices may be smart devices, network-enabled devices, devices with connections to WiFi / 3G / 4G / 5G / 6G / Zigbee / Bluetooth / NFC / UMTS / 3GPP / GSM / EDGE / TDMA / FDMA / CDMA / WCDMA / TD-SCDMA / ad hoc networks / other networks, smart speakers, smart watches, smart clocks, smart appliances, smart machines, smart appliances, smart tools, smart vehicles, Internet of Things (IoT) devices, Internet-enabled devices, computers, portable computers, tablets, and other devices. Nearby devices and / or at least one processor associated with a wireless receiver, a wireless transmitter, another wireless receiver, another wireless transmitter, and / or a cloud server (in the cloud) may determine an initial STI for an object. Two or more of them may jointly determine initial spatio-temporal information. Two or more of them may share intermediate information in determining the initial STI (e.g., initial location).

[0160] In one example, a wireless transmitter (e.g., a Type 1 device or a tracker bot) moves with an object. The wireless transmitter can send a signal to a wireless receiver (e.g., a Type 2 device or an Origin Register) or determine the object's initial STI (e.g., initial location). The wireless transmitter can also send a signal and / or another signal to another wireless receiver (e.g., another Type 2 device or another Origin Register) to monitor the object's movement (spatio-temporal information). The wireless receiver can also receive a signal and / or another signal from the wireless transmitter and / or another wireless transmitter to monitor the object's movement. The location of the wireless receiver and / or another wireless receiver can be known. In another example, a wireless receiver (e.g., a Type 2 device or a tracker bot) can move with the object. The wireless receiver can receive a signal transmitted from the wireless transmitter (e.g., a Type 1 device or an Origin Register) to determine the object's initial spatio-temporal information (e.g., initial location). The wireless receiver may also receive a signal and / or another signal from another wireless transmitter (e.g., another Type 1 device or another origin register) for monitoring the current movement (e.g., space-time information) of the object. The wireless transmitter may also transmit a signal and / or another signal to the wireless receiver and / or another wireless receiver (e.g., another Type 2 device or another tracker bot) for monitoring the movement of the object. The location of the wireless transmitter and / or another wireless transmitter may be known.

[0161] Locations include sensing areas, rooms, houses, offices, property, workspaces, corridors, lifts, lift wells, escalators, elevators, sewers, ventilation systems, stairs, assembly areas, ducts, air ducts, pipes, enclosed spaces, enclosed structures, semi-enclosed structures, enclosed areas with at least one wall, plants, machines, engines, structures, structures with wood, structures with glass, structures with metal, structures with walls, structures with doors, structures with gaps, structures with reflective surfaces, structures with liquids, buildings, rooftops, stores, factories, assembly lines, homes, etc. Hotel rooms, museums, classrooms, schools, universities, government buildings, warehouses, garages, malls, airports, train stations, bus terminals, hubs, transportation hubs, cargo terminals, government buildings, public facilities, schools, universities, entertainment venues, recreational facilities, hospitals, pediatric / neonatal wards, nursing homes, elderly care facilities, community centers, stadiums, playgrounds, fields, basketball courts, tennis courts, soccer stadiums, baseball fields, gymnasiums, garages, shopping marts, knolls, supermarkets, manufacturing facilities, parking facilities Facilities, construction sites, mining facilities, transportation facilities, highways, roads, valleys, forests, trees, terrain, landscapes, caves, patios, land, roads, amusement parks, urban areas, rural areas, suburban areas, metropolitan areas, gardens, squares, plazas, music halls, downtown facilities, open facilities, semi-open facilities, closed areas, train platforms, train stations, distribution centers, warehouses, shops, distribution centers, storage facilities, underground spaces, spatial (e.g. above ground, space) facilities, floating facilities, caves, tunnel facilities, indoor facilities, outdoor facilities, outdoor facilities with some walls / doors / reflective barriers Areas such as facilities, open facilities, semi-open facilities, automobiles, trucks, buses, vans, containers, ships / boats, submarines, trains, trams, airplanes, vehicles, mobile platforms, caves, tunnels, pipes, channels, metropolitan areas, downtown areas with relatively tall buildings, valleys, wells, ducts, pathways, gas lines, oil pipes, water pipes, interconnecting pathways / arrays / roads / tubes / cavities / caves / pipe-like structures / voids / fluid spaces, human bodies, animal bodies, body cavities, organs, bones, teeth, soft tissue, hard tissue, rigid tissue, non-hard tissue, blood / body fluid ducts, wind ducts, air ducts, burrows, etc. The location may be an indoor space, an outdoor space, and the location may include both inside and outside spaces.For example, a location can include both the inside and outside of a building. For example, a location can be a building with one or more floors, and part of the building can be underground. The shape of the building can be, for example, round, square, rectangular, triangular, or irregular. These are just examples. The present disclosure can be used to detect events in other types of locations or spaces.

[0162] The wireless transmitter (e.g., a Type 1 device) and / or wireless receiver (e.g., a Type 2 device) may be embedded in a portable device (e.g., a module or a device having a module) that may travel with the object (e.g., in a previous travel and / or a current travel). The portable device may be communicatively coupled to the object using a wired connection (e.g., via USB, micro USB, Firewire, HDMI, serial port, parallel port, and other connectors) and / or a connection (e.g., Bluetooth, Bluetooth Low Energy (BLE), WiFi, LTE, NFC, ZigBee). The portable device may be a lightweight device. The portable device may be powered by batteries, rechargeable batteries, and / or AC power. The portable device may be very small (e.g., on the sub-millimeter and / or sub-centimeter scale) and / or small (e.g., coin-sized, card-sized, pocket-sized, or larger). The portable device may be large, bulky, and / or heavy installed machinery. Portable devices include WiFi hotspots, access points, Mobile WiFi (MiFi), dongles with USB / micro USB / Firewire / other connectors, smartphones, portable computers, computers, tablets, smart devices, Internet of Things (IoT) devices, WiFi enabled devices, LTE enabled devices, smart watches, smart glass, smart mirrors, smart antennas, smart batteries, smart lights, smart pens, smart rings, smart doors, smart windows, smart clocks, smart batteries, smart wallets, smart belts, smart handbags, smart cloth / garments, smart ornaments, smart packaging, smart paper / books / magazines / posters / printed materials / signage / displays / illuminated systems / lighting systems, smart keys / tools, smart bracelets / chains / necklaces / clothes / accessories, smart pads / cushions, smart tiles / blocks / bricks / building materials / other materials,Smart trash can / waste container, smart food carriage / storage, smart ball / racket, smart chair / sofa / bed, smart shoes / footwear / carpet / mat / shoe rack, smart gloves / handwear / ring / handwear, smart hat / cap / cosmetics / sticker / tattoo, smart mirror, smart toy, smart pill, smart cookware, smart bottle / food container, smart tool, smart device, IoT device, WiFi enabled device, network enabled device, 3G / 4G / 5G / 6G enabled device, UMTS device, 3GPP device, GSM device, EDGE device, TDMA device, FDMA device, CDMA device, WCDMA device, TD-SCDMA device, embedded device, embeddable device, air conditioner, refrigerator, heater, furnace, furniture, oven, cooking device , TV / Set Top Box (STB) / DVD Player / Audio Player / Video Player / Remote Control, Hi-Fi, Audio Device, Speaker, Lamp / Light, Wall, Door, Window, Roof, Tile / Roofing Shingle / Structure / Attic Structure / Device / Feature / Installation / Fixture, Lawn Mower / Garden Equipment / Tools / Machine Tools / Garage Tools, Garbage Can / Container, 20ft / 40ft Container, Storage Container, Factory / Production / Manufacturing Equipment, Repair Tool, Fluid Container, Machine, Installed Machine, Vehicle, Cart, Wagon, Warehouse Vehicle, Automobile, Bicycle, Motorcycle, Boat, Watercraft, Airplane, Basket / Box / Bag / Bucket / Container, Smart Plate / Cup / Bowl / Pot / Mat / Utensil / Kitchenware / Kitchen Accessories / Cabinet / Table / Chair / Tile / Light / Water Pipe / Faucet / Gas Range / Oven / Dishwasher / , etc. Portable devices may have batteries that may be replaceable, non-replaceable, rechargeable, and / or non-rechargeable. Portable devices may be charged wirelessly. The portable device may be a smart payment card. The portable device may be a payment card used in parking lots, highways, entertainment parks, or other locations / facilities requiring payment. The portable device may have an identity (ID) / identifier, as described above.

[0163] Events can be monitored based on the TSCI. Events can be object-related events such as an object (e.g., a person and / or a patient) falling, rolling, hesitating, resting, impact (e.g., a person hitting a punching bag, door, window, bed, chair, table, desk, cabinet, box, another person, animal, bird, flying, table, chair, ball, bowling ball, tennis ball, football, soccer ball, baseball, basketball, volleyball), two-body action (e.g., leaving a balloon, catching a fish, molding clay, writing a paper, person typing into a computer), moving a car in a garage, a person carrying a smartphone and walking around an airport / mall / government office / office / etc, autonomous mobile objects / machines moving around (e.g., vacuum cleaner, utility vehicle, car, drone, self-driving car), etc.

[0164] The tasks or wireless smart sensing tasks are object detection, presence detection, proximity detection, object recognition, activity recognition, object verification, object counting, daily activity monitoring, health monitor, vital signs monitoring, health condition monitoring, baby monitoring, elderly monitoring, sleep monitoring, sleep stage monitoring, gait monitoring, motion monitoring, tool detection, tool recognition, tool verification, patient detection, patient monitoring, patient verification, machine detection, machine verification, human detection, human recognition, human verification, baby detection, baby recognition, baby verification, human breathing detection, human breathing recognition, human breathing estimation, human breathing verification, human heart rate detection, human heart rate recognition, human heart rate estimation, human heart rate verification, fall detection, fall recognition, fall estimation, fall verification, emotion detection, emotion recognition, emotion estimation, emotion verification, motion detection, motion degree estimation, Motion recognition, motion estimation, motion verification, cyclic motion detection, cyclic motion estimation, cyclic motion verification, repetitive motion detection, cyclic motion recognition, repetitive motion estimation, repetitive motion verification, static motion recognition, static motion detection, static motion estimation, static motion verification, cyclostationary motion detection, cyclostationary motion recognition, cyclostationary motion estimation, cyclostationary motion verification, transient motion detection, transient motion recognition, transient motion estimation, transient motion verification, trend detection, trend recognition, trend estimation, trend verification, breathing detection, breathing recognition, breathing estimation, human biometric detection, human biometric recognition, human biometric estimation, human biometric verification, environmental informatics detection, environmental informatics recognition, environmental informatics estimation, environmental informatics verification, gait detection, gait recognition, gait estimation, gait verification, gesture detection, gesture recognition, gesture estimation, gesture verification, machine learning, supervised learning, unsupervised learning, semi-supervised learning, clustering, feature extraction, feature training, principal component analysis, eigenvalue decomposition, frequency decomposition, time decomposition,Time-Frequency Decomposition, Function Decomposition, Other Decompositions, Training, Discriminative Training, Supervised Training, Unsupervised Training, Semi-Supervised Training, Neural Networks, Sudden Motion Detection, Fall Detection, Hazard Detection, Life-Threat Detection, Regular Motion Detection, Stationary Motion Detection, Cyclostationary Motion Detection, Intrusion Detection, Suspicious Motion Detection, Security, Safety Monitoring, Navigation, Guidance, Map-Based Processing, Map-Based Correction, Model-Based Processing / Correction, Irregularity Detection, Localization, Room Sensing, Tracking, Multiple Object Tracking, Indoor Tracking, Indoor Positioning, Indoor Navigation, Energy Management, Power Transmission, Wireless Power Transmission, Object Counting, Car Tracking in Parking Garages, Device / System Activation (e.g., Security Systems, Access Systems, Alarms, Sirens, Speakers, Televisions, Entertainment Systems, Cameras, Heating / Air Conditioning (HVAC) Systems, Ventilation Systems, Lighting Systems, Gaming Systems, Coffee Machines, Cooking Appliances, Cleaning Equipment, Housekeeping Equipment), Geometric Estimation , augmented reality, wireless communication, data communication, signal broadcasting, networking, coordination, management, encryption, protection, cloud computing, other processing, and / or other tasks, which may be performed by a Type 1 device, a Type 2 device, another Type 1 device, another Type 2 device, a nearby device, a local server (such as a hub device), an edge server, a cloud server, and / or another device.This task can be based on TSCI between any pair of Type 1 and Type 2 devices. A Type 2 device can also be a Type 1 device, and vice versa. A Type 2 device can fulfill the role (e.g., functionality) of a Type 1 device temporarily, continuously, sporadically, simultaneously, and / or concurrently, and / or vice versa. The first portion of tasks may include at least one of pre-processing, processing, signal conditioning, signal processing, post-processing, sporadic / continuous / concurrent / simultaneous / dynamic / adaptive / on-demand / as needed processing, calibration, noise removal, feature extraction, coding, encryption, transformation, mapping, motion detection, motion estimation, motion change detection, motion pattern detection, motion pattern estimation, motion pattern recognition, vital sign detection, vital sign estimation, vital sign recognition, periodic motion detection, periodic motion estimation, repetitive motion detection / estimation, breathing rate detection, breathing rate estimation, breathing pattern detection, breathing pattern estimation, breathing pattern recognition, heart rate detection, heart rate estimation, cardiac pattern detection, cardiac pattern estimation, cardiac pattern recognition, gesture detection, gesture estimation, gesture recognition, velocity detection, velocity estimation, object location, object tracking, navigation, acceleration estimation, acceleration detection, fall detection, change detection, intruder (and / or tampering) detection, baby detection, baby monitoring, patient monitoring, object recognition, wireless power transfer, and / or wireless charging.

[0165] The second part of the task may be a smart home task, a smart office task, a smart building task, a smart factory task (e.g., manufacturing with a machine or assembly line), a smart Internet of Things (IoT) task, a smart system task, a smart home operation, a smart office operation, a smart building operation, a smart manufacturing operation (e.g., movement of supplies / parts / raw materials to a machine / assembly line), an IoT operation, a smart system operation, turning on lights, turning off lights, controlling light in at least one of a room, area, and / or location, playing a sound clip, playing a sound clip in at least one of a room, area, and / or location, playing at least one sound clip of welcome, greeting, farewell, a first message, and / or a second message related to the first part of the task, turning on an appliance, turning off an appliance, The control may include at least one of controlling appliances in at least one of the rooms, areas, and / or locations; turning on an electrical system; turning off an electrical system; controlling an electrical system in at least one of the rooms, areas, and / or locations; turning on a security system; turning off a security system; controlling a security system in at least one of the rooms, areas, and / or locations; turning on a mechanical system; turning off a mechanical system; controlling a mechanical system in at least one of the rooms, areas, and / or locations; and / or controlling at least one of an air conditioning system, a heating system, a ventilation system, a lighting system, a heater, a stove, an entertainment system, a door, a fence, a window, a garage, a computer system, a networked device, a networked system, a home appliance, an office appliance, a lighting device, a robot (e.g., a robotic arm), a smart vehicle, a smart machine, an assembly line, a smart device, an Internet of Things (IoT) device, a smart home device, and / or a smart office device.

[0166] The tasks are to detect when the user comes home, detect when the user leaves, detect when the user moves from one room to another, detect when a window / door / garage door / blinds / curtains / panels / solar panels / sunshades are controlled / locked / unlocked / opened / closed / partially opened, detect pets, detect / monitor when the user is doing something (e.g. sleeping on the sofa, sleeping in the bedroom, running on the treadmill, cooking, sitting on the sofa, watching TV, eating in the kitchen, eating in the dining room, going up and down stairs, going out / coming back, in the bathroom), monitor / detect the location of the user / pet and do something automatically when detected (e.g. sending a message, notifying / reporting), do something to the user when a user is detected, turn on / off / dim the lights, turn on / off the music / radio / home entertainment system, turn on / off the TV / may include turning on / off / adjusting / controlling hi-fi / set-top box (STB) / home entertainment system / smart speaker / smart device, turning on / off / adjusting air conditioning system, turning on / off / adjusting ventilation system, turning on / off / adjusting heating system, adjusting / controlling curtains / light shades, turning on / off / starting computer, turning on / off / preheating / controlling coffee machine / hot water kettle, turning on / off / preheating / controlling cooker / oven / microwave / other cooking appliance, checking / adjusting temperature, checking weather forecast, checking phone message box, checking email, checking system, controlling / adjusting system, checking / controlling / arming / disarming security system / baby monitor, checking / controlling refrigerator, reporting (e.g. through speaker such as Google Home, Amazon Echo, via web page / email / messaging system / notification system).

[0167] For example, if a user arrives at home in their vehicle, the tasks may automatically detect that the user or their vehicle is approaching, open the garage door upon detection, turn on the driveway / garage lights as the user approaches the garage, turn on the air conditioner / heater / fan, etc. As the user enters the house, the tasks may automatically turn on the entrance lights, turn off the driveway / garage lights, play a greeting message to welcome the user, turn on music, turn on the radio and tune it to the user's favorite radio news channel, open the curtains / blinds, monitor the user's mood, adjust the lighting and sound environment according to the user's mood or current / impending events on the user's daily calendar (e.g., romantic lighting and music because the user is going to have dinner with his / her girlfriend in an hour), microwave the food the user prepared that morning, perform a diagnostic check of all systems in the house, check the weather forecast for tomorrow's tasks, check news of interest to the user, and update the user's calendar, to-do list, etc. Check reminders, check phone answering systems, messaging systems, email, communicate verbal reports using dialogue systems / speech synthesis, remind the user of their mother's birthday (e.g., using audible tools such as speakers, hi-fi, speech synthesis, sound, voice, music, song, sound field, background sound field, dialogue system, etc.; using visual tools such as TV / entertainment system / computer / notebook / smartpad / display / light / color / brightness / pattern, using tactile tools / virtual reality tools / gestures / tools, using smart devices / appliances / materials / furniture / fixtures, using web tools / servers / hub devices / cloud servers / fog servers / edge servers / home networks / mesh networks, using messaging tools / notification tools / communication tools / scheduling tools / email, using user interfaces / GUIs, using scents / smells / aromas / tastes, using neural tools / neural system tools, or a combination), generate reports, provide reports (e.g., using reminding tools as described above).A task may be to proactively start an air conditioner / heater / ventilation system or proactively adjust a smart thermostat temperature setting. When a user moves from the front door to the living room, the tasks may be to turn on the living room lights, open the living room curtains, open the windows, turn off the front door light behind the user, turn on the TV and set-top box, turn on the set-top box, set the TV to the user's preferred channel, adjust the appliances according to the user's preferences and conditions / states (e.g., adjust the lighting, select / play music to create a romantic atmosphere), etc.

[0168] Another example could be: When a user wakes up in the morning, the task could be to detect the user moving around in the bedroom, open the blinds / curtains, open the windows, turn off the alarm clock, adjust the room temperature profile from a night temperature profile to a day temperature profile, turn on the bedroom lights, turn on the toilet light as the user approaches the bathroom, check the radio or streaming channels, play the morning news, turn on the coffee machine, preheat water, turn off the security system, etc. When the user walks from the bedroom to the kitchen, the task could be to turn on the kitchen and hallway lights, turn off the bedroom and toilet lights, move music / messages / reminders from the bedroom to the kitchen, turn on the kitchen TV, change the TV to the morning news channel, lower the kitchen blinds, open the kitchen window to let in fresh air, unlock the back door so the user can check the back yard, adjust the kitchen temperature setting, etc. Another example could be: When the user leaves home for work, the tasks may be to detect the user's departure, say goodbye and / or have a nice day, open / close the garage door, turn on / off the garage and driveway lights, turn off / dim to save energy (only if the user fails), close / lock all windows / doors (only if the user fails), turn off appliances (especially the stove, oven, microwave), turn on / arm the home security system to protect the home against intruders, adjust the air conditioning / heating / ventilation system to an "away from home" profile to save energy, send alerts / reports / updates to the user's smartphone, etc.

[0169] The motions are classified as no motion, rest motion, motionless motion, motion, change of location / position, deterministic motion, transient motion, falling motion, repetitive motion, periodic motion, pseudo-periodic motion, periodic / repetitive motion related to breathing, periodic / repetitive motion related to heartbeat, periodic / repetitive motion related to living organisms, periodic / repetitive motion related to machines, periodic / repetitive motion related to man-made objects, periodic / repetitive motion related to nature, complex motion related to transient and periodic elements, repetitive motion, non-deterministic motion, stochastic motion, chaotic motion, random motion, complex motion with non-deterministic and deterministic elements, stationary random motion, pseudo-stationary random motion, cyclostationary random motion, non-stationary random motion, non-stationary random motion with a periodic autocorrelation function (ACF), random motion with a periodic ACF over time. motion, pseudo-stationary random motion over time, random motion where the instantaneous ACF has a pseudo-periodic / repetitive component over time, machine motion, mechanical motion, vehicle motion, drone motion, air-related motion, wind-related motion, weather-related motion, water-related motion, fluid-related motion, ground-related motion, change in magnetic properties, subsurface motion, earthquake motion, plant motion, animal motion, animal motion, human motion, normal motion, abnormal motion, dangerous motion, warning motion, suspicious motion, rain, fire, flood, tsunami, explosion, collision, imminent collision, human motion, head motion, face motion, eye motion, oral motion, tongue motion, neck motion, finger motion, hand motion, arm motion, shoulder motion, body motion, chest motion, abdominal motion, hip motion, leg motion, foot motion, body joint motion, knee motion, elbow motion, upper body motion, lower body motion, skin motion, subcutaneous motion, subcutaneous tissue motion. The motion may include at least one of blood vessel movement, venous movement, organ movement, heart movement, lung movement, stomach movement, intestinal movement, bowel movement, eating movement, breathing movement, facial expression, eye expression, mouth expression, vocal movement, singing movement, eating movement, gesture, hand gesture, arm gesture, keystroke, typing stroke, user interface gesture, man-machine interaction, gait, dance movement, coordinated movement, and / or coordinated body movement.

[0170] The heterogeneous ICs of the Type 1 device and / or any Type 2 receiver may include a low noise amplifier (LNA), a power amplifier, a transmit-receive switch, a media access controller, a baseband radio, a 2.4 GHz radio, a 3.65 GHz radio, a 4.9 GHz radio, a 5 GHz radio, a 5.9 GHz radio, a sub-6 GHz radio, a 60 GHz radio, a sub-60 GHz radio, and / or another radio. The heterogeneous IC may include a processor, a memory communicatively coupled to the processor, and a set of instructions stored in the memory for execution by the processor. The IC and / or any processor may include at least one of a general purpose processor, a special purpose processor, a microprocessor, a multiprocessor, a multi-core processor, a parallel processor, a CISC processor, a RISC processor, a microcontroller, a central processing unit (CPU), a graphical processor unit (GPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), an embedded processor (e.g., ARM), a logic circuit, another programmable logic device, discrete logic, and / or a combination. Heterogeneous ICs are used in broadband networks, wireless networks, mobile networks, mesh networks, cellular networks, wireless local area networks (WLANs), wide area networks (WANs), metropolitan area networks (MANs), WLAN standards, WiFi, LTE, LTE-A, LTE-U, 802.11 standards, 802.11a, 802.11b, 802.11g, 802.11n, 802.11ac, 802.11af, 802.11ah, 802.11ax, 802.11ay, mesh networking standards 802.16, 3G, 3.5G, 4G, Beyond 4G, 4.5G, 6G, 7G, 8G, 9G, UMTS, 3GPP, GSM, EDGE, TDMA, FDMA, CDMA, WCDMA, TD-SCDMA, Bluetooth, Bluetooth Low-Energy It may support Bluetooth Low Energy (BLE), NFC, Zigbee, WiMax, and other wireless network protocols.

[0171] The processor may include a general-purpose processor, a special-purpose processor, a microprocessor, a microcontroller, an embedded processor, a digital signal processor, a central processing unit, a graphical processing unit (GPU), a multiprocessor, a multi-core processor, and / or a processor with graphics capabilities, and / or a combination thereof. The memory may be volatile, non-volatile, random-access memory (RAM), read-only memory (ROM), programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), a hard disk, flash memory, CD-ROM, DVD-ROM, magnetic storage, optical storage, organic storage, a storage system, a storage network, network storage, cloud storage, edge storage, local storage, external storage, internal storage, or any other form of non-transitory storage medium known in the art. The set of instructions (machine-executable code) corresponding to the method steps may be directly embodied in hardware, software, firmware, or a combination thereof. The set of instructions may be embedded, pre-loaded, loaded at boot time, loaded on-the-fly, loaded on-demand, pre-installed, installed, and / or downloaded.

[0172] The presentation may be visual (e.g., using a combination of sights, graphics, text, symbols, color, shade, video, animation, sound, voice, audio, etc.), graphical (e.g., using GUI, animation, video), textual (e.g., web page with text, messages, animated text), symbolic (e.g., emojis, signs, hand gestures), or mechanical (e.g., vibration, actuator movement, haptics, etc.) presentation.

[0173] Basic calculation

[0174] The computational workload associated with this method is shared among the processor, the type 1 heterogeneous wireless device, the type 2 heterogeneous wireless device, a local server (eg, a hub device), a cloud server, and other processors.

[0175] Operations, pre-processing, processing, and / or post-processing may be applied to the data (e.g., TSCI, autocorrelation, TSCI features). Operations may include pre-processing, processing, and / or post-processing. Pre-processing, processing, and / or post-processing may be operations. Operations may include pre-processing, processing, post-processing, scaling, calculating confidence coefficients, calculating line-of-sight (LOS) quantities, non-LOS calculations, and the like. Computing non-linear (NLOS) quantities, computing quantities including line-of-sight and non-linear (NLOS), computing quantities for a single link (e.g., path, communication path, link between a transmitting antenna and a receiving antenna), computing quantities including multiple links, computing functions of operands, filtering, linear filtering, nonlinear filtering, folding, grouping, energy computation, low-pass filtering, band-pass filtering, high-pass filtering, median filtering, rank filtering, quartile filtering, percentile filtering, finite impulse response (FIR) filtering, infinite impulse response (IIR) filtering, moving average (MA) filtering, autoregressive (AR) filtering, autoregressive moving average (ARMA) filtering, selective filtering, adaptive filtering, interpolation, decimation, subsampling, upsampling, resampling, time correction, time-based correction, phase correction, magnitude correction, phase cleaning, amplitude cleaning, matched filtering, enhancement, restoration, noise removal, smoothing, signal conditioning, enhancement, restoration, linear transform, nonlinear transform, inverse transform, frequency transformation, inverse frequency transformation, Fourier transform (FT), discrete-time FT (DTFT), Discrete FT (DFT), Fast FT (FFT), Wavelet Transform, Laplace Transform, Hilbert Transform, Hadamard Transform, Trigonometric Transform, Sine Transform, Cosine Transform, DCT, Power of 2 Transform, Sparse Transform, Graph Based Transform, Graph Signal Processing, Fast Transform, Transform Combined with Zero Padding, Cyclic Padding, Padding, Zero Padding, Feature Extraction, Decomposition, Projection, Orthogonal Projection, Non-Orthogonal Projection, Overprojection (oecomlee)ojecion), eigendecomposition, singular value decomposition (SVD), principle component analysis (ICA), independent component analysis (ICA), grouping, sorting, thresholding, soft thresholding, hard thresholding, clipping, soft clipping, first derivative, second derivative, higher derivative, convolution, multiplication, division, addition, subtraction, integration, maximization, minimization, least squared error, recursive least squares, constrained least squares, batch least squares, least absolute deviation, least mean squared deviation, least absolute deviation, local maximization, local minimization, cost function optimization, neural networks, recognition, labeling, training, class filtering, machine learning, supervised learning, unsupervised learning, semi-supervised learning, comparison with other TSCI, similarity score calculation, quantization, vector quantization, matching pursuit, compression, encryption, encoding, storage, transmission, normalization, time normalization, frequency domain normalization, classification, clustering, labeling, tagging, learning, detection, estimation, learning network, mapping, remapping, expansion, storage, search, transmission, reception, representation, combining, merging, splitting, tracking, monitoring, matched filtering, Kalman filtering, particle filtering, interpolation, extrapolation, histogram estimation The operations may include importance sampling, Monte Carlo sampling, compressed sensing, representation, merging, combining, dividing, scrambling, error protection, forward error correction, do nothing, time-varying processing, adjusted averaging, weighted averaging, arithmetic mean, geometric mean, harmonic mean, averaging over selected frequencies, averaging over antenna links, logical operations, permutation, combination, sorting, AND, OR, XOR, union, intersection, vector addition, vector subtraction, vector multiplication, vector division, inverse, norm, distance, and / or another operation. The operations may be pre-processing, processing, and / or post-processing. Operations may be applied jointly on multiple time series or functions.

[0176] Functions (e.g., functions of operands) can be scalar functions, vector functions, discrete functions, continuous functions, polynomial functions, properties, characteristics, magnitude, phase, exponential functions, logarithmic functions, trigonometric functions, transcendental functions, boolean functions, linear functions, algebraic functions, nonlinear functions, piecewise functions, real functions, complex functions, vector-valued functions, inverse functions, derivative functions, integral functions, circular functions, functions of other functions, one-to-one functions, one-to-many functions, many-to-one functions, many-to-many functions, zero crossings, absolute functions, index functions, mean, mode, median, range, statistics, histogram, variance, standard deviation, measure of change, expansion, dispersion, deviation, divergence, range, interquartile range, total deviation, absolute deviation, arithmetic mean, geometric mean, harmonic mean, trimmed mean, percentile, square, cube, square root, power, sine, cosine, tangent, cotangent elliptic functions, parabolic functions, hyperbolic functions, game functions, zeta functions, absolute value, threshold, limit functions, floor functions, rounding functions, sign functions, quantization, piecewise constant functions, composite functions, functions of functions, time functions processed by operations (e.g. filtering), probabilistic functions, stochastic functions, deterministic functions, periodic functions, iterative functions, transformations, frequency transforms, inverse frequency transforms, discrete time transforms, Laplace transforms, Hilbert transforms, sine transforms, cosine transforms, trigonometric transforms, wavelet transforms, integer transforms, power of two transforms, sparse transforms, projections, decompositions, principal component analysis (PCA), neural networks, feature extraction, moving functions, functions for moving windows of adjacent items in a time series, filtering functions, convolutions, mean functions, histograms, variance / standard deviation functions, statistical functions, short-time transforms, discrete transforms, Discrete Fourier Transform, Discrete Cosine Transform, Discrete Sine Transform, Hadamard Transform, Eigenvalue Decomposition, Eigenvalue, Singular Value Decomposition (SVD), Singular Value, Orthogonal Decomposition, Matching Pursuit, Sparse Transform, Arbitrary Decomposition, Graph Based Processing, Graph Based Transform, Graph Signal Processing, Classification, Class / Group / Category Identification, Labeling, Learning, Machine Learning, Detection, Estimation, Feature Extraction, Learning Networks, Feature Extraction, Noise Reduction, Signal Enhancement, Coding, Encryption, Mapping, Remapping, Vector Quantization, Low-Pass Filtering, High-Pass Filtering, Band-Pass Filtering, Matched Filtering, Kalman Filtering, Pre-Processing, Post-Processing, Particle Filtering, FIR Filtering, IIR Filtering, Autoregressive (AR) Filtering, Adaptive Filtering, First Derivative, Higher-Order Derivative, Integration, Zero Crossing, Smoothing, Median Filtering, Mode Filtering, Sampling, Random Sampling, Resampling Function, Downsampling, Downconverting, Upsampling, Upconverting, Interpolation, Extrapolation, Importance Sampling, Monte Carlo Sampling, Compressed Sensing, Statistics, These may include short-term statistics, long-term statistics, autocorrelation functions, cross-correlation functions, moment-generating functions, time averages, weighted averages, special functions, Bessel functions, error functions, complementary error functions, beta functions, gamma functions, integral functions, Gaussian functions, Poisson functions, and the like.

[0177] The steps (or each step) of this disclosure may employ machine learning, training, discriminative training, deep learning, neural networks, continuous time processing, distributed computing, distributed storage, and acceleration using GPUs / DSPs / coprocessors / multi-cores / multi-processing.

[0178] The frequency transform may include a Fourier transform, a Laplace transform, a Hadamard transform, a Hilbert transform, a sine transform, a cosine transform, a trigonometric transform, a wavelet transform, an integer transform, a power of two transform, zero-padding and combinations of transforms, a power Fourier transform with zero-padding, and / or another transform. Fast and / or approximate versions of the transforms may be performed. The transforms may be performed using floating-point and / or fixed-point arithmetic.

[0179] The inverse frequency transform may include an inverse Fourier transform, an inverse Laplace transform, an inverse Hadamard transform, an inverse Hilbert transform, an inverse sine transform, an inverse cosine transform, an inverse trigonometric transform, an inverse wavelet transform, an inverse integer transform, an inverse power of two transform, a combination of zero padding and transforms, an inverse Fourier transform with zero padding, and / or another transform. Fast and / or approximate versions of the transform may be performed. The transform may be performed using floating-point and / or fixed-point arithmetic.

[0180] Quantities / features can be calculated from TSCI. Quantities include: movement, location, map coordinates, height, speed, acceleration, movement angle, rotation, dimension, volume, time trend, one-time pattern, recurring pattern, evolving pattern, time pattern, mutually exclusive pattern, associated / correlated pattern, cause-effect, short-term / long-term correlation, tendency, slope, preference, statistics, typical behavior, atypical behavior, time trend, time profile, periodic movement, repetitive movement, repetition, tendency, change, sudden change, gradual change, frequency, transient, breathing, gait, behavior, event, suspicious event, dangerous event, warning event, warning, belief, proximity, collision, power, signal, signal power, signal strength, signal volume, received signal strength indicator (RSSI), signal amplitude, signal phase, signal frequency component, signal frequency band component, channel state information (CSI), map, time, frequency, time-period The statistics may include at least one of wavenumber, decomposition, orthogonal decomposition, non-orthogonal decomposition, tracking, respiration, palpitations, statistical parameters, cardiopulmonary statistics / analysis (e.g., output response), daily activity statistics / analysis, chronic disease statistics / analysis, medical statistics / analysis, early (or instantaneous or simultaneous or delayed) indicator / suggestion / sign / marker / verifier / detection / symptom / condition / state, biometric, baby, patient, machine, device, temperature, vehicle, parking lot, location, lift, elevator, space, road, fluid flow, home, room, office, house, building, warehouse, storage, system, ventilation, fan, pipe, duct, people, human, car, boat, truck, plane, drone, downtown, crowd, impulse event, cyclostationary, environment, vibration, material, surface, 3D, 2D, local, global, presence, and / or other measurable quantity / variable.

[0181] Sliding Window Algorithm

[0182] The sliding time window may have a time-varying window width. It may be smaller initially to allow for rapid acquisition and may increase over time to a steady-state size. The steady-state size may be related to the monitored frequency, repetitive motion, transient motion, and / or STI. Even in the steady state, the window size may be adaptively (and / or dynamically) changed (e.g., adjusted, varied, modified) based on battery life, power consumption, available computing power, changes in the volume of interest, the nature of the monitored motion, etc.

[0183] The time shift between two sliding time windows at adjacent time instances can be constant / variable / locally adaptive / dynamically adjusted over time. If a shorter time shift is used, any monitoring updates can be more frequent, which can be used for rapidly changing conditions, object movements, and / or objects. A longer time shift can be used for slower conditions, object movements, and / or objects.

[0184] The window width / size and / or time shift may be changed (e.g., adjusted, altered, modified) according to user requests / selections. The time shift may be changed automatically (e.g., as controlled by a processor / computer / server / hub device / cloud server) and / or adaptively (and / or dynamically).

[0185] At least one characteristic (e.g., a feature value or a feature point) of the function (e.g., an autocorrelation function, an autocovariance function, a cross-correlation function, a cross-covariance function, a power spectral density, a time function, a frequency domain function, a frequency transform) may be determined (e.g., by the object tracking server, a processor, a Type 1 heterogeneous device, a Type 2 heterogeneous device, and / or another device). At least one characteristic of the function may include a maximum, a minimum, an extremum, a local maximum, a local minimum, a local extremum, a local extremum with a positive time offset, a first local extremum with a positive time offset, an nth local extremum with a positive time offset, a first local extremum with a negative time offset, a bounded maximum, a bounded minimum, a bounded extremum, a significant maximum, a significant minimum, a significant extremum, a gradient, a derivative, a higher order derivative, a maximum gradient, a minimum gradient, a local maximum gradient, a local maximum gradient with a positive time offset, a local minimum gradient, a bounded maximum gradient, a bounded minimum gradient, a maximum higher order derivative, a minimum higher order derivative, a bounded higher order derivative, a zero crossing, a zero crossing with a positive time offset, an nth zero crossing with a positive time offset, a zero crossing with a negative time offset, an nth zero crossing with a negative time offset, a bounded zero crossing, a zero crossing of a gradient, a zero crossing of a gradient of a higher order derivative, and / or other characteristics. At least one argument of the function associated with at least one characteristic of the function may be identified. A quantity (eg, spatial-temporal information of an object) may be determined based on at least one argument of the function.

[0186] Characteristics (e.g., characteristics of an object's movement at a location) can include instantaneous characteristics, short-term characteristics, recurring characteristics, recursive characteristics, history, incremental characteristics, change characteristics, deviation characteristics, phase, amplitude, degree, time characteristics, frequency characteristics, time-frequency characteristics, decomposition characteristics, orthogonal decomposition characteristics, non-orthogonal decomposition characteristics, deterministic characteristics, probability characteristics, stochastic characteristics, autocorrelation function (ACF), mean, variance, standard deviation, measure of change, spread, variance, deviation, divergence, range, interquartile range, total variation, absolute deviation, total deviation, statistics, duration, timing, trend, periodic characteristics, recurring characteristics, long-term characteristics, historical characteristics, average characteristics, recent characteristics, past characteristics, future characteristics, predicted characteristics, position, distance, height, speed, direction, velocity, acceleration, change in acceleration, angle, angular speed, change in angular velocity, of an object The change in the angular acceleration includes at least one of: angular acceleration, change in angular acceleration, orientation of the object, angle of rotation, deformation of the object, shape of the object, change in shape of the object, change in size of the object, change in structure of the object, and / or change in properties of the object.

[0187] At least one local maximum and at least one local minimum of the function may be identified. At least one local signal-to-noise ratio-like (SNR-like) parameter may be calculated for each pair of adjacent local maximums and minima. The SNR-like parameter may be a function (e.g., linear, logarithmic, exponential, monotonic) of the fraction of the amount (e.g., power, magnitude) of the local maximum over the same amount of the local minimum. It may also be a function of the difference between the amount of the local maximum and the same amount of the local minimum. Significant local peaks may be identified or selected. Each significant local peak may be a local maximum with an SNR-like parameter greater than a threshold T1 and / or a local maximum with an amplitude greater than a threshold T2. At least one local minimum and at least one local minimum in the frequency domain may be identified / calculated using a persistence-based approach.

[0188] A set of selected significant local peaks may be selected from the set of identified significant local peaks based on a selection criterion (e.g., quality criteria, signal quality state). An object characteristic / STI may be calculated based on the set of selected significant local peaks and frequency values ​​associated with the set of selected significant local peaks. In one example, the selection criterion may always correspond to selecting the strongest peak in the range. The strongest peak may be selected, but non-selected peaks may still be significant (or even strong).

[0189] Unselected significant peaks may be saved and / or monitored as "reserved" peaks for use in future selections in future sliding time windows. As an example, there may be a particular peak (at a particular frequency) that appears consistently over time. Initially, even though it is significant, it may not be selected (because other peaks may become stronger). However, at a later time, the peak may become stronger and more dominant and may be selected. If it is "selected," it may be back-traced in time and deemed "selected" at an earlier time that was significant but not selected. In such a case, the back-traced peak may replace a previously selected peak at an earlier time. The replaced peak may be a relatively weak peak or a peak that appears isolated in time (i.e., appears only briefly in time).

[0190] In other instances, the selection criteria may not correspond to selecting the strongest peak in the range, but instead may consider not only the "strength" of the peak, but also the "trace" of peaks that may have occurred in the past, especially peaks that have been identified for a long time.

[0191] For example, if a finite state machine (FSM) is used, it may select the peak based on the state of the FSM. The decision threshold may be adaptively (and / or dynamically) calculated based on the state of the FSM.

[0192] The similarity score and / or component similarity score may be calculated (e.g., by a server (e.g., a hub device), a processor, a Type 1 device, a Type 2 device, a local server, a cloud server, and / or another device) based on a pair of temporally adjacent CIs of a TSCI. The pairs may be obtained from the same sliding window or two different sliding windows. The similarity score may also be based on a pair of temporally adjacent or less-adjacent CIs from two different TSCIs. The similarity score and / or component similarity score may be / include time reversal resonating strength (TRRS), correlation, cross-correlation, autocorrelation, correlation indicator, covariance, cross-covariance, autocovariance, dot product of two vectors, distance score, norm, metric, quality metric, signal quality condition, statistical property, discrimination score, neural network, deep learning network, machine learning, training, discrimination, weighted average, preprocessing, denoising, signal conditioning, filtering, time correction, time alignment, phase offset compensation, transform, component-wise operation, feature extraction, finite state machine, and / or another score. Property and / or STI may be determined / calculated based on the similarity score.

[0193] Any threshold may be predetermined, adaptively (and / or dynamically) determined, and / or determined by a finite state machine. Adaptive determination may be based on time, space, location, antenna, path, link, condition, battery life, remaining battery capacity, available power, available computational resources, available network bandwidth, etc.

[0194] A threshold to be applied to a test statistic to distinguish between two events (or two conditions, or two situations, or two states), A and B, may be determined. Data (e.g., CI, channel state information (CSI), power parameters) may be collected under A and / or under B in a training scenario. The trial statistic may be calculated based on the data. The distribution of the trial statistic under A may be compared to the distribution of the trial statistic under B (a reference distribution), and the threshold may be selected according to some criteria. The criteria may include maximum likelihood estimation (ML), maximum a posteriori probability (MAP), discriminative training, minimum type-1 error for a given type-2 error, minimum type-2 error for a given type-1 error, and / or other criteria (e.g., quality metric, signal quality condition). The threshold may be adjusted to achieve different sensitivities to A, B, and / or other events / conditions / situations / states. The threshold adjustment may be automatic, semi-automatic, and / or manual. Threshold adjustments may be applied once, occasionally, frequently, periodically, repeatedly, occasionally, sporadically, and / or on-demand. Threshold adjustments may be adaptive (and / or dynamically adjusted). Threshold adjustments may depend on objects, object movement / location / orientation / motion, object characteristics / STI / size / characteristics / traits / habits / behavior, location, at / at / of location, features / fixtures / furniture / barriers / materials / machines / creatures / objects / boundaries / surfaces / media, map, map constraints (or environmental model), events / states / scenes / conditions, time, timing, duration, current state, past history, user, and / or personal preference, etc.

[0195] A stopping criterion (or skip or bypass or blocking or pausing or passing or rejecting criterion) for an iterative algorithm may be that the change in the current parameter (e.g., offset value) in the update in the iteration is less than a threshold. The threshold may be 0.5, 1, 1.5, 2, or another number. The threshold may be adaptive (and / or dynamically adjusted). It may change as the iteration progresses. With respect to the offset value, the adaptive threshold may be determined based on the task, the initial specific value, the current time offset value, the regression window, the regression analysis, the regression function, the regression error, the convexity of the regression function, and / or the number of iterations.

[0196] The local extrema may be determined as a corresponding extremum of the regression function in the regression window. The local extrema may be determined based on a set of time offset values ​​and a set of associated regression function values ​​within the regression window. Each of the set of associated regression function values ​​associated with the set of time offset values ​​may be within a range from the corresponding extremum of the regression function in the regression window.

[0197] Searching for local extrema, robust search, minimization, optimization, statistical optimization, dual optimization, constraint optimization, convex optimization, global optimization, local optimization, energy minimization, linear regression, quadratic regression, higher order regression, linear programming, nonlinear programming, stochastic programming, combinatorial optimization, constraint programming, constraint satisfaction, computation of variations, optimal control, dynamic programming, mathematical programming, multiobjective optimization, multimodal optimization, disjunctive programming, space mapping, infinite dimensional optimization, heuristics, metaheuristics, convex programming, semidefinite programming, cone programming, second order cone programming, integer programming, quadratic programming, fractional programming, numerical analysis, simplex algorithm, iterative methods, gradient descent, subgradient methods, coordinate gradient methods, conjugate gradient methods, Newton's algorithm, sequential quadratic programming, interior point methods, elliptic methods, reduced gradient methods, quasi-Newton methods, simultaneous perturbation stochastic approximation, interpolation, pattern search methods, line search, non-differential optimization, genetic algorithms, evolutionary algorithms, dynamic relaxation The search for local extrema may involve an objective function, a loss function, a cost function, a utility function, a fitness function, an energy function, and / or an energy function.

[0198] The regression may be performed using a regression function to fit the sampled data (e.g., CIs, CI features, components of CIs) or another function (e.g., an autocorrelation function) in a regression window. The length of the regression window and / or the position of the regression window may be varied in at least one iteration. The regression function may be a linear function, a quadratic function, a cubic function, a polynomial function, and / or another function.

[0199] The regression analysis may minimize at least one of the following: error, aggregate error, component error, error in a projected domain, error in a selected axis, error in a selected orthogonal axis, absolute error, squared error, absolute deviation, squared deviation, higher-order error (e.g., third-order, fourth-order), robust error (e.g., squared error versus absolute error for smaller magnitude errors and for larger errors, or a first type of error for smaller magnitude errors and a second type of error for larger magnitude errors), another error, a weighted sum (or weighted average) of absolute / squared errors (e.g., in the case of a wireless transmitter with multiple antennas and a wireless receiver with multiple antennas, each pair of transmitter antenna and receiver antenna forms a link), mean absolute error, mean squared error, mean absolute deviation, and / or mean squared deviation. Errors associated with different links may have different weights. One possibility is that some links and / or some components with greater noise or lower signal quality metrics may have smaller or larger weights. (weighted sum of squared errors, weighted sum of higher-order errors, weighted sum of robust errors, weighted sum of alternative errors, absolute cost, squared cost, higher-order cost, robust cost, alternative cost, weighted sum of absolute costs, weighted sum of squared costs, weighted sum of higher-order costs, weighted sum of robust costs, and / or weighted sum of alternative costs).

[0200] The determined regression error may be an absolute error, a squared error, a higher order error, a robust error, a further error, a weighted sum of absolute errors, a weighted sum of squared errors, a weighted sum of higher order errors, a weighted sum of robust errors, and / or a weighted sum of further errors.

[0201] The time offset associated with the maximum regression error (or minimum regression error) of the regression function for a particular function within the regression window may be the updated current time offset for the iteration.

[0202] The local extrema can be searched for based on a quantity including the difference between two different errors (e.g., the difference between an absolute error and a squared error), each of which can include an absolute error, a squared error, a higher-order error, a robust error, another error, a weighted sum of absolute errors, a weighted sum of squared errors, a weighted sum of higher-order errors, a weighted sum of robust errors, and / or a weighted sum of another error.

[0203] The quantity may be compared to reference data or a reference distribution, such as an F-distribution, a central F-distribution, another statistical distribution, a threshold, a threshold associated with a probability / histogram, a threshold associated with a probability / histogram of finding a false peak, a threshold associated with an F-distribution, a threshold associated with a central F-distribution, and / or a threshold associated with another statistical distribution.

[0204] The regression window may be determined based on at least one of: a movement of the object (e.g., a change in location / position), a quantity related to the object, at least one characteristic and / or STI of the object related to the movement of the object, an estimated location of a local extremum, a noise characteristic, an estimated noise characteristic, a signal quality metric, an F-distribution, a central F-distribution, another statistical distribution, a threshold, a preset threshold, a threshold related to a probability / histogram, a threshold related to a desired probability, a threshold related to the probability of finding a false peak, a threshold related to an F-distribution, a threshold related to a central F-distribution, a threshold related to another statistical distribution, a condition that the quantity at the window center is a maximum within the regression window, a condition that only one of the local extrema of a particular function for a particular value exists for the first time within the regression window, another regression window, and / or other conditions.

[0205] The width of the regression window can be determined based on the particular local extrema being searched for, including a first local maximum, a second local maximum, a higher-order local maximum, a first local maximum with a positive time offset, a second local maximum with a positive time offset, a higher-order local maximum with a positive time offset, a first local maximum with a negative time offset, a second local maximum with a negative time offset, a second local maximum with a negative time offset, a higher-order local maximum with a negative time offset, a first local minimum, a second local minimum, a higher-order local minimum, a first local minimum with a positive time offset, The local minimum may include a second local minimum with a positive time offset, a higher-order local minimum with a positive time offset, a first local minimum with a negative time offset, a second local minimum with a negative time offset, a higher-order local minimum with a negative time offset, a first local extremum, a second local extremum, a higher-order local extremum, a first local extremum with a positive time offset, a second local extremum with a positive time offset, a first local extremum with a negative time offset, a second local extremum with a negative time offset, and a higher-order extremum with a negative and / or negative time offset.

[0206] The current parameters (e.g., time offset value) may be initialized based on a target value, a target profile, a trend, a past trend, a current trend, a target speed, a speed profile, a target speed profile, a past speed trend, an object's motion or movement (e.g., a change in location / position), at least one characteristic and / or STI of the object associated with the object's motion, a position quantity of the object, an initial speed of the object associated with the object's motion, a predefined value, an initial width and duration of the regression window, a value based on the signal's carrier frequency, a value based on the signal's subcarrier frequency, a signal's bandwidth, an antenna's aggregate value associated with the channel, noise characteristics, a signal h metric, and / or an adaptive (and / or dynamically adjusted) value. The current time offset may be at the center, left, right, and / or another fixed relative position of the regression window.

[0207] In the presentation, the information may be displayed together with a map (or environmental model) of the location. The information may include location, zone, area, region, coverage area, corrected location, approximate location, location wrt a map of the location, location wrt a segmented location, direction, route, route wrt a map and / or segmentation, trace (e.g., location within a time window such as the last 5 seconds, or the last 10 seconds, the time window duration may be adjusted adaptively (and / or dynamically), and the time window duration may be adjusted adaptively (and / or dynamically) with respect to speed, acceleration), route history, approximate area / zone along the route, history / summary of past locations, history of past locations of interest, frequently visited areas, customer traffic, flock fabric, flock behavior, flock control information, speed, acceleration, movement statistics, breathing rate, heart rate, presence / absence of movement, presence or absence of people or pets or objects, presence or absence of vital signs, gesture control (controlling a device using gestures), location-based gesture control, information for location-based operations, objects of interest (e.g., pets, people, self-guided machines / devices, vehicles, drones, cars, boats, bicycles, etc.). Identity (ID) or identifier of a vehicle (e.g., car, unmanned vehicle, machine with fan, air conditioner, TV, machine with moving parts), user identification (e.g., person), user information, position / speed / acceleration / direction / movement / gesture / gesture control / movement trace, user ID or identifier, user activity, user status, user sleep / rest characteristics, user emotional state, user vital signs, location environment information, location weather information, earthquake, explosion, storm, rain, fire, temperature, collision, impact, vibration, event, door opening event, door closing event, window opening The events may include a window closing event, a fall event, a burning event, an icy event, a water-related event, a wind-related event, an air movement event, an accident event, a quasi-periodic event (e.g., running on a treadmill, hopping, skipping rope, somersaults, etc.), a repeating event, a swarming event, a vehicle event, a user gesture (e.g., hand gestures, arm gestures, foot gestures, leg gestures, body gestures, head gestures, face gestures, mouth gestures, eye gestures, etc.).

[0208] Location may be two-dimensional (e.g., using two-dimensional coordinates), three-dimensional (e.g., using three-dimensional coordinates). Location may be relative (e.g., with respect to a map or environmental model) or relational (e.g., halfway between point A and point B, around a corner, upstairs, on a table, on the ceiling, on the floor, on a couch, close to point A, a distance R from point A, within a radius of R from point A, etc.). Location may be expressed in Cartesian coordinates, polar coordinates, and / or another representation.

[0209] Information (e.g., location) may be marked with at least one symbol. The symbol may change over time. The symbol may flash and / or pulsate with or without changing color / intensity. The size may change over time. The orientation of the symbol may change over time. The symbol may be a number reflecting an instantaneous quantity (e.g., user's vital signs / respiratory rate / heart rate / gesture / status / condition / action / movement, temperature, network traffic, network connectivity, device / machine status, remaining device power, device state, etc.). The rate of change, size, orientation, color, intensity, and / or symbol may reflect the respective movement. Information may be presented visually and / or verbally explained (e.g., using pre-recorded audio or speech synthesis). Information may be written in text. Information may also be presented in a mechanical manner (e.g., animated gadgets, moving parts).

[0210] The user interface (UI) device may be a smartphone (e.g., iPhone, Android phone), a tablet (e.g., iPad), a laptop (e.g., notebook computer), a personal computer (PC), a device with a graphic user interface (GUI), a smart speaker, a device with voice / sound / speaker capabilities, a virtual reality (VR) device, an augmented reality (AR) device, a smart car, an in-car display, a voice assistant, an in-car voice assistant, etc.

[0211] A map (or environmental model) may be two-dimensional, three-dimensional, and / or higher dimensional. (e.g., a time-varying 2D / 3D map / environment model). Walls, windows, doors, entrances, exits, and restricted areas may be marked on the map or model. A map may include a floor plan of a facility. A map or model may have one or more layers (overlays). A map / model may be a maintenance map / model including water pipes, gas pipes, wiring, cable runs, air ducts, crawl spaces, ceiling layouts, and / or basement layouts. A location may be segmented / subdivided / regionalized / grouped into multiple zones / areas / geographical regions / sectors / sections / territories / districts / administrative areas / sites / neighborhoods / areas / stretches / open spaces, such as bedrooms, living rooms, storage rooms, walkways, kitchens, dining rooms, foyers, garages, first floors, second floors, restrooms, offices, conference rooms, reception areas, various office areas, various warehouse areas, various facility areas, etc. The segments / areas / regions may be presented on the map / model. Different regions may be color coded. Different regions may be presented with characteristics (e.g., color, brightness, color intensity, texture, animation, blinking, blink rate, etc.) Logical segmentation of locations may be performed using at least one heterogeneous Type 2 device, or server (e.g., hub device), or cloud server, etc.

[0212] Here is an example of the disclosed system, device, and method. Stephan and his family want to install the disclosed wireless motion detection system to detect movement in their 2,000-square-foot, two-story townhouse in Seattle, Washington. Because his house is two stories, Stephan decides to use one Type 2 device (named A) and two Type 1 devices (named B and C) on the first floor. The first floor is centered around three rooms: the kitchen, dining room, and living room, with the dining room in the middle and arranged in a straight line. The kitchen and living room are on opposite sides of the house. He places a Type 2 device (A) in the dining room, one Type 1 device (B) in the kitchen, and another Type 1 device (C) in the living room. With this device installation, he specifically uses the motion detection system to partition the first floor into three zones: the dining room, the living room, and the kitchen. When motion is detected by the AB pair and the AC pair, the system analyzes the motion information and associates the motion with one of three zones.

[0213] When Stefan and his family go away for the weekend (e.g., going camping for a long weekend), Stefan turns on the motion detection system using a mobile phone app (e.g., Android phone app or iPhone app). When the system detects motion, an alert signal is sent to Stefan (e.g., SMS text message, email, push message to the mobile phone app, etc.). If Stefan pays a monthly fee (e.g., $10 / month), a service company (e.g., a security company) receives the alert signal through a wired network (e.g., broadband) or wireless network (e.g., home WiFi, LTE, 3G, 2.5G, etc.) and performs security procedures for Stefan (e.g., calling him to check the problem, sending someone to check the house, contacting the police on Stefan's behalf, etc.). Stefan loves his elderly mother and is concerned about her well-being when he is home alone. When his mother is home alone while the rest of the family is out (e.g., going to work, shopping, or on vacation), Stephan uses his mobile app to turn on the motion detection system to ensure his mother is okay. He then uses the mobile app to monitor his mother's movements around the house. When Stephan uses the mobile app to see his mother moving around the house between three areas, according to her daily routine, Stephan knows that his mother is okay. Stephan is grateful that the motion detection system can help him monitor his mother's well-being while he is away from home.

[0214] On a typical day, his mother wakes up around 7:00 AM. She plans to make breakfast in the kitchen in about 20 minutes. She then eats breakfast in the dining room for about 30 minutes. Then, she does her daily exercise in the living room before sitting on the sofa and watching her favorite TV show. The motion detection system allows Stephan to see the timing of movements in each of three areas of the house. When the movements fit into her daily routine, Stephan knows roughly that his mother should be doing well. However, if the movement pattern seems abnormal (e.g., no movement until 10:00 AM, staying in the kitchen too long, remaining motionless for long periods of time, etc.), Stephan suspects something is wrong and calls his mother to check on her. Stephan may even ask someone (e.g., family member, neighbor, paid staff member, friend, social worker, service provider) to check on his mother.

[0215] Occasionally, Stephen feels the need to reposition his Type 2 device. He simply unplugs the device from its original AC power plug and plugs it into another AC power plug. He is pleased that the wireless motion detection system is plug-and-play, and repositioning does not affect the system's operation. Once powered on, it works immediately.

[0216] On another occasion, Stephan is so convinced that our wireless motion detection system can indeed detect motion with very high accuracy and very low alerts that he can actually use the mobile app to monitor motion on the first floor. He decides to install a similar configuration (i.e., one Type 2 device and two Type 1 devices) on the second floor to monitor the bedrooms on the second floor. Again, he finds that system setup is extremely easy, as he simply needs to plug the Type 2 and Type 1 devices into AC power plugs on the second floor. No special installation is required. He can then use the same mobile app to monitor motion on the first and second floors. Each Type 2 device on the first / second floors can interact with all Type 1 devices on both the first and second floors. Stephan is happy to see that as he doubles his investment in Type 1 and Type 2 devices, he has more than double the capacity of his combined system.

[0217] According to various embodiments, each CI (CI) may include at least one of channel state information (CSI), frequency domain CSI, a frequency representation of CSI, frequency domain CSI associated with at least one subband, time domain CSI, intra-domain CSI, a channel response, a channel response estimate, a channel impulse response (CIR), a channel frequency response (CFR), channel characteristics, a channel filter response, CSI of a wireless multipath channel, information of a wireless multipath channel, a timestamp, auxiliary information, data, metadata, user data, account data, access data, security data, session data, status data, management data, family data, identity (ID), identifier, device data, network data, proximity data, environmental data, real-time data, sensor data, stored data, encrypted data, compressed data, protected data, and / or another CI. In one embodiment, the disclosed system includes hardware components (e.g., a wireless transmitter / receiver with an antenna, analog circuitry, a power supply, a processor, a memory) and corresponding software components. According to various embodiments of the present disclosure, the disclosed system includes a Bot (referred to as a Type 1 device) and an Origin (referred to as a Type 2 device) for vital signs detection and monitoring, each device comprising a transceiver, a processor, and a memory.

[0218] The disclosed system can be applied in many ways. In one example, a Type 1 device (transmitter) may be a small WiFi-enabled device placed on a table. It may also be a WiFi-enabled television (TV), set-top box (STB), smart speaker (e.g., Amazon Echo), smart refrigerator, smart microwave, mesh network router, mesh network satellite, smartphone, computer, tablet, smart plug, etc. In one example, a Type 2 (receiver) may be a WiFi-enabled device placed on a table. It may also be a WiFi-enabled television (TV), set-top box (STB), smart speaker (e.g., Amazon Echo), smart refrigerator, smart microwave, mesh network router, mesh network satellite, smartphone, computer, tablet, smart plug, etc. Type 1 and Type 2 devices may be placed in / near a conference room to count people. Type 1 and Type 2 devices may be a health monitoring system for the elderly to monitor daily activities and any signs of symptoms (e.g., dementia, Alzheimer's disease). Type 1 and Type 2 devices may be used in an infant monitor to monitor the vital signs (breathing) of living infants. Type 1 and Type 2 devices can be placed in bedrooms to monitor sleep quality and any sleep apnea. Type 1 and Type 2 devices can be placed in automobiles to monitor passenger and driver health, detect driver sleep, and detect any babies left in the car. Type 1 and Type 2 devices can be used in logistics to prevent human trafficking by monitoring people hidden in trucks and containers. Type 1 and Type 2 devices can be deployed by emergency services in disaster areas to search for victims trapped in rubble. Type 1 and Type 2 devices can be placed in an area to detect the breathing of any intruders. Non-wearable wireless respiratory monitoring has many applications.

[0219] The hardware modules may be configured to include Type 1 transceivers and / or Type 2 transceivers and may be sold / used under variable brands to design, build, and sell final commercial products. The products using the disclosed systems and / or methods may be home / office security products, sleep monitoring products, WiFi products, mesh products, TVs, STBs, entertainment systems, HiFi, speakers, home appliances, lamps, stoves, ovens, microwaves, tables, chairs, beds, shelves, tools, appliances, torches, vacuum cleaners, smoke detectors, sofas, pianos, fans, doors, windows, door / window handles, locks, smoke detection equipment, car accessories, computing devices, office supplies, air conditioners, heaters, pipes, connectors, surveillance cameras, access points, computer equipment, mobile devices, LTE devices, 3G / 4G / 5G / 6G devices, UMTS devices, 3GPP devices, GSM devices, EDGE devices, TDMA devices, FDMA devices, CDMA devices, WCDMA devices, TD-SCDMA devices, gaming devices, eyeglasses, glass panels, VR goggles, necklaces, watches, waistbands, belts, wallets, pens, hats, clothing, implantable devices, tags, parking tickets, smartphones, etc.

[0220] The summarization may include: analysis, output response, selected time window, sub-sampling, transformation, and / or projection. The presentation may include presenting at least one of a month / week / day view, simplified / detailed view, cross-sectional view, small / large form factor view, color-coded view, comparison view, summary view, video, web view, audio announcement, and another presentation related to the cyclical / recurring characteristics of the recurring motion.

[0221] A Type 1 / Type 2 device is any device that includes an antenna, a device with an antenna, a device with a housing (e.g., for a radio, antenna, data / signal processing unit, radio IC, circuitry), a device that interfaces / attaches / connects / links to another device / system / computer / phone / network / data aggregator, a device with a user interface (UI) / graphical UI / display, a device with a wireless transceiver, a device with a wireless transmitter, a device with a wireless receiver, an Internet of Things (IoT) device, a device with a wireless network, a device with both wired and wireless network capabilities, a device with a wireless integrated circuit (IC), a Wi-Fi device, a device with a Wi-Fi chip (e.g., compliant with 802.11a / b / g / n / ac / ax standards), a Wi-Fi access point (AP), a Wi-Fi client, a WiFi router, a Wi-Fi repeater, a WiFi hub, a WiFi mesh network router / hub / AP, a wireless mesh network router, an ad-hoc network device, a wireless mesh network device, a mobile device (e.g., 2G / 2.5G / 3G / 3G).5G / 4G / LTE / 5G / 6G / 7G, UMTS, 3GPP, GSM, EDGE, TDMA, FDMA, CDMA, WCDMA, TD-SCDMA), cellular device, base station, mobile network base station, mobile network hub, mobile network compatible device, LTE device, device with LTE module, mobile module (e.g., circuit board with mobile-enabling chip (IC) such as Wi-Fi chip, LTE chip, BLE chip), device with mobile module, smartphone, companion device for smartphone (e.g., dongle, attachment, plug-in), dedicated device, plug-in device, AC-powered device, battery-powered device, device with processor / memory / instruction set, smart device / gadget / Item: A watch, stationery, pen, user interface, paper, mat, camera, television (TV), set-top box, microphone, speaker, refrigerator, oven, machine, phone, wallet, furniture, door, window, ceiling, floor, wall, table, chair, bed, nightstand, air conditioner, heater, pipe, duct, cable, carpet, decoration, gadget, USB device, plug, dongle, lamp / light, tile, ornament, bottle, vehicle, automobile, AGV, drone, robot, laptop, tablet, computer, hard disk, network card, equipment, racket, ball, shoe, wearable device, clothing, eyeglasses, hat, necklace, food, pill, small device that moves within a living being's body (e.g., blood vessels, lymph, digestive system), and / or another device. The Type 1 device and / or Type 2 device may be communicatively coupled to the Internet, another device that accesses the Internet (e.g., a smartphone), a cloud server (e.g., a hub device), an edge server, a local server, and / or storage. Type 1 and / or Type 2 devices may operate under local control, may be controlled by another device via a wired or wireless connection, may operate automatically, or may be controlled by a central system located remotely (e.g., away from the home).

[0222] In one embodiment, a Type-B device may be a transceiver that may perform as both an origin (Type-2 device, Rx device) and a bot (Type-1 device, Tx device), i.e., a Type-B device may be both a Type-1 (Tx) device and a Type-2 (Rx) device (e.g., simultaneously or alternately), such as a mesh device, mesh router, etc. In one embodiment, a Type-A device may be a transceiver that may function only as a bot (Tx device), i.e., it may be only a Type-1 device or only a Tx, such as a simple IoT device. It may have the functionality of an origin (Type-2 device, Rx device), but in some embodiment, it functions only as a bot. All Type-A and Type-B devices form a tree structure. The root may be a Type-B device that has access to a network (e.g., the Internet). For example, it may be connected to broadband service via a wired connection (e.g., Ethernet, cable modem, ADSL / HDSL modem) or a wireless connection (e.g., LTE, 3G / 4G / 5G, WiFi, Bluetooth, microwave link, satellite link, etc.). In one embodiment, all Type A devices are leaf nodes. Each Type B device may be a root node, a non-leaf node, or a leaf node.

[0223] The Type 1 device (transmitter, or Tx) and Type 2 device (receiver, or Rx) may be on the same device (e.g., RF chip / IC), or simply on the same device. The device may operate in high frequency bands such as 28 GHz, 60 GHz, 77 GHz, etc. The RF chip may have dedicated Tx antennas (e.g., 32 antennas) and dedicated Rx antennas (e.g., another 32 antennas).

[0224] One Tx antenna can transmit a radio signal (e.g., a series of probe signals, perhaps at 100 Hz). Alternatively, all Tx antennas can be used to transmit radio signals with beamforming (at Tx), so that the radio signals are focused in a particular direction (e.g., for energy efficiency, or to boost the signal-to-noise ratio in that direction, or low-power operation when "scanning" in that direction, or low-power operation when an object is known to be in that direction).

[0225] The radio signal hits an object (e.g., a living human lying on a bed 4 feet away from the Tx / Rx antennas, breathing and heartbeat) within a location (e.g., a room). Object movement (e.g., lung movement according to breathing rate, or blood vessel movement according to heartbeat) can affect / modulate the radio signal. All Rx antennas can be used to receive the radio signal.

[0226] Beamforming (at the Rx and / or Tx) may be applied (digitally) to "scan" different directions. Many directions may be scanned or monitored simultaneously. Along with beamforming, a "sector" (e.g., direction, orientation, azimuth, bearing, zone, region, segment) may be defined relative to the Type 2 device (e.g., relative to the center position of the antenna array). For each probing signal (e.g., pulse, ACK, control packet, etc.), channel information or CI (e.g., channel impulse response / CIR, CSI, CFR) is obtained / calculated (e.g., from the RF chip) for each sector. For respiration detection, CIR can be collected over a sliding window (e.g., 30 seconds; a 100 Hz ringing / probing rate could have 3000 CIRs over 30 seconds).

[0227] A CIR can have many taps (e.g., N1 components / tap). Each tap may be associated with a time lag, or time-of-fright (e.g., the time it takes to hit and back a person 4 feet away). When breathing in a certain direction at a certain distance (e.g., 4 feet), one can find the CIR in the "certain direction" and then find the tap corresponding to the "certain distance." Respiration rate and heart rate can then be calculated from that tap of that CIR.

[0228] Each tap within a sliding window (e.g., a 30-second window of "component time series") can be considered a time function (e.g., a "tap function," "component time series"). Each tap function can be examined in search of strong periodic behavior (e.g., corresponding perhaps to breathing in the range 10 bpm to 40 bpm).

[0229] A Type 1 device and / or a Type 2 device can have external connections / links and / or internal connections / links. An external connection (e.g., connection 1110) can be associated with 2G / 2.5G / 3G / 3.5G / 4G / LTE / 5G / 6G / 7G / NBIoT, UWB, WiMax, Zigbee, 802.16, etc. The internal connections (e.g., 1114A and 1114B, 1116, 1118, 1120) can be associated with WiFi, IEEE802.11 standards, 802.11a / b / g / n / ac / ad / af / ag / ah / ai / aj / aq / ax / ay, Bluetooth 1.0 / 1.1 / 1.2 / 2.0 / 2.1 / 3.0 / 4.0 / 4.1 / 4.2 / 5, BLE, mesh networking, and IEEE802.16 / 1 / 1a / 1b / 2 / 2a / a / b / c / d / e / f / g / h / i / j / k / l / m / n / o / p / standards.

[0230] Type 1 devices and / or Type 2 devices are powered by batteries (e.g., AA batteries, AAA batteries, coin cell batteries, button cell batteries, small batteries, battery banks, power banks, car batteries, hybrid batteries, vehicle batteries, container batteries, non-rechargeable batteries, rechargeable batteries, NiCd batteries, NiMH batteries, lithium ion batteries, zinc carbon batteries, zinc chloride batteries, lead acid batteries, alkaline batteries, batteries with wireless chargers, smart batteries, solar batteries, boat batteries, plain batteries, other batteries, temporary energy storage devices, capacitors, flywheels).

[0231] Any device may be powered by DC or direct current (e.g., from batteries, generators, power converters, solar panels, rectifiers, DC-DC converters as described above, at various voltages such as 1.2V, 1.5V, 3V, 5V, 6V, 9V, 12V, 24V, 40V, 42V, 48V, 110V, 220V, 380V, etc.) and therefore may have a DC connector or connectors with at least one pin for DC power.

[0232] Any device may be powered by AC or alternating current (e.g., from a domestic wall outlet, a transformer, an inverter, shore power, or various voltages such as 100V, 110V, 120V, 100-127V, 200V, 220V, 230V, 240V, 220-240V, 100-240V, 250V, 380V, 50Hz, 60Hz, etc.) and therefore may have an AC connector or connectors with at least one pin for AC power. Type 1 devices and / or Type 2 devices may be located (e.g., installed, positioned, moved) within or outside a location.

[0233] For example, in a vehicle (e.g., an automobile, truck, lorry, bus, specialty vehicle, tractor, excavator, drilling machine, teleporter, bulldozer, crane, forklift, electric vehicle, AGV, emergency vehicle, cargo, freight car, trailer, container, boat, ferry, ship, submarine, aircraft, airship, lift, monorail, train, electric railcar, rail car, rail car, etc.), the Type 1 device and / or Type 2 device may be an embedded device embedded in the vehicle or an add-on device (e.g., an aftermarket device) plugged into a port within the vehicle (e.g., an OBD port / socket, a USB port / socket, an accessory port / socket, a 12V auxiliary power outlet, and / or a 12V cigarette lighter port / socket).

[0234] For example, one device (e.g., a Type 2 device) may be plugged into a 12V cigarette lighter / accessory port or an OBD port or a USB port (e.g., of a car / truck / vehicle), and the other device (e.g., a Type 1 device) may be plugged into a 12V cigarette lighter / accessory port or an OBD port or a USB port (e.g., of a car / truck / vehicle). The OBD port and / or USB port may provide power, signaling, and / or networking (of the car / truck / vehicle). The two devices may jointly monitor passengers, including children / babies, in the vehicle. They may be used to count passengers, recognize the driver, and detect the presence of passengers in specific seats / positions within the vehicle.

[0235] In another embodiment, one device may be plugged into a 12V cigarette lighter / accessory port or OBD port or USB port of a car / truck / vehicle, while the other device may be plugged into a 12V cigarette lighter / accessory port or OBD port or USB port of another car / truck / vehicle.

[0236] In another example, many devices of the same Type A (e.g., Type 1 or Type 2) may exist in many heterogeneous vehicles / portable devices / smart gadgets (e.g., automated guided vehicles / AGVs, shopping / luggage / mobile carts, parking tickets, golf carts, bicycles, smartphones, tablets, cameras, recording devices, smartwatches, roller skates, shoes, jackets, goggles, hats, eyewear, wearables, Segways, scooters, baggage tags, cleaning machines, vacuum cleaners, pet tags / collars / wearables / implants), each of which may be plugged into the vehicle's 12V accessory port / OBD port / USB port or embedded in the vehicle. There may also be one or more other Type B devices (e.g., if A is Type 2, then B is Type 1, and if A is Type 1, then B is Type 2) installed in locations such as gas stations, streetlights, street corners, tunnels, multi-story parking lots, and scattered locations covering a large area such as factories, stadiums, train stations, shopping malls, and construction sites. Type A devices can be located, tracked, or monitored based on TSCI.

[0237] The area / location may not have local connectivity such as broadband service, WiFi, etc. Type 1 and / or Type 2 devices may be portable. Type 1 and / or Type 2 devices may support plug and play.

[0238] Pairwise wireless links can be established between many pairs of devices, forming a tree structure. In each pair (and associated link), a device (the second device) may be a non-leaf (Type B). The other device (the first device) may be a leaf (Type A or Type B) or a non-leaf (Type B). In the link, the first device acts as a bot (Type 1 device or transmitting device) to transmit a wireless signal (e.g., a probe signal) to the second device over a wireless multipath channel. The second device can act as an origin (Type 2 device or Rx device) to receive the wireless signal, obtain TSCI, and calculate a "link analysis" based on the TSCI.

[0239] While not all computing devices have cameras or speakers, almost all of them include one or more wireless modules. While keystroke recognition can be performed using 2.4 GHz / 5 GHz WiFi radios, these WiFi-based approaches are limited by narrow bandwidths, long wavelengths, and a limited number of antennas. For example, limited by 20 MHz / 40 MHz bandwidths, these systems' range resolution can be several meters, where reflected signals from all targets and the background environment are superimposed and difficult to separate. Furthermore, WiFi-based approaches require cumbersome data-driven training to achieve reasonable recognition accuracy, are unable to recognize multiple simultaneous keystrokes due to mixed signals, and are typically trained on a single, fixed keyboard layout.

[0240] In this disclosure, a wireless tracking system or virtual keyboard system (referred to as "mmKey") is designed, according to various embodiments, based on millimeter wave (mmWave) radio to wirelessly track keystrokes on a virtual keyboard. Without any additional hardware, mmKey can transform any flat surface, such as printed paper or painted areas, into an interactive typing medium. Compared to conventional approaches, mmKey simultaneously enables the distinct features of keystroke support and user-defined keyboard layouts. In some embodiments, without requiring any training, mmKey achieves all of these features in a universal virtual keyboard system by capturing mmWave signals reflected from moving fingers and using a novel pipeline of signal processing. As a result, mmKey is environment-independent and position-independent, so it works anywhere and can easily adapt to various keyboards, such as computer keyboards, piano keyboards, phone keypads, or other user-customized layouts, at zero cost.

[0241] mmKey overcomes multiple challenges to provide a practical system for general-purpose millimeter-wave radio. First, it is important to design a robust motion detector that can capture minute movements on the keyboard before keystroke recognition is possible. To address this challenge, mmKey first applies z-score anomaly detection to the amplitude difference of the channel impulse response (CIR) to sense signal fluctuations and infer the presence of motion. Because the carrier frequency is high, the signal decays rapidly over propagation distance, so the threshold for keystroke detection should adapt to the distance. By referencing the empty CIR measured in the absence of a target, an adaptive z-score detector can be designed. Multiple antennas and different ranges can be further utilized to improve robustness.

[0242] Second, because keystrokes involve not only finger movements but also palm and potentially arm movements, it is difficult to distinguish between keystrokes (finger movements) and other movements. Furthermore, there are also irrelevant reflections from background objects that are mixed in with the keystroke movements. To overcome this challenge, we first devise a novel motion filter by exploiting the sensitivity of CIR phase along with the difference in the spatial distribution of dynamic signals between keystrokes and other types of movements. Then, adaptive background cancellation can be utilized to extract only dynamic reflections by tracking changes in CIR.

[0243] Third, despite the large number of antennas in mmWave devices, spatial resolution can be physically limited by the small effective aperture of the receive antenna array. In some embodiments, on-chip analog beamforming can provide 15° angular resolution on an exemplary experimental device with an array size of 1.8 cm × 1.8 cm, which is insufficient to localize and recognize keystrokes, especially when the key size is very small or simultaneous keystrokes are close to each other. To increase spatial resolution, some mmKey embodiments run the MUltiple Signal Classification (MUSIC) algorithm on the received CIR, enabling accurate localization of keystrokes. Only initial finger localization can reveal the location of the movement relative to the device. A low-effort, single calibration phase can be used during initial setup to determine the keys pressed by the user, which can be as simple as three key presses, allowing the MUSIC-estimated locations to be mapped to the corresponding keys on the keyboard.

[0244] In some embodiments, the mmKey system can be implemented on a general-purpose 60 GHz 802.11ad / ay network chipset with an additional array attached to enable radar-like operation and report CIR. mmKey's performance has been verified through extensive experiments involving 10 volunteers in different positions in both home and office environments on three different virtual keyboards, including a computer keyboard, piano keyboard, and telephone keypad. According to some embodiments, experimental results demonstrate remarkable accuracy rates of over 95% for single-keystroke scenarios and over 90% for multiple simultaneous keystrokes. Furthermore, by feeding mmKey's output into a commercially available text collection tool, significant word recognition accuracy rates of over 97% can be achieved relative to natural typing on a printed computer keyboard. With its superior performance, mmKey provides a universal virtual keyboard for computers, mobile devices, wearables, and IoT devices equipped with millimeter-wave radio.

[0245] FIG. 1A illustrates an exemplary device configuration for an mmKey system according to some embodiments of the present disclosure. As shown in FIG. 1A, the mmKey system includes a device, a transmitter (Tx) antenna array 110, and a receiver (Rx) antenna array 120. In some embodiments, the device operates in the 60 GHz frequency band with a bandwidth of 3.52 GHz. In some embodiments, the transmitter (Tx) and receiver (Rx) arrays each have 32 antennas assembled in a 6x6 configuration. To extract the CIR, the Tx 110 can transmit bursts consisting of a group of 32 pulses, which are sequentially received by different Rx antennas 120 after being reflected by the surrounding environment, as shown in FIG. 1A.

[0246] In some embodiments, the transmitter 110 is a bot as described above, and the receiver 120 is an origin as described above. While in FIG. 1A the transmitter 110 and receiver 120 are physically coupled to each other, in other embodiments they may be separated into separate devices. In some embodiments, the device including the Tx 110 and Rx 120 functions like a radar, and keystrokes are tracked when the surface 101 faces the radar.

[0247] In some embodiments, as shown in FIG. 1C, the duration of each pulse is Tp=10 μs and the duration of each burst is Tb=100 ms. By examining the difference in the time of arrival (ToA) corresponding to the distinct propagation paths, the system can distinguish between reflectors located at different ranges. In some embodiments, the 3.52 GHz bandwidth on the experimental device provides a time resolution of 0.28 ns, which means that two paths with a delay difference greater than 0.28 ns can be distinguished, corresponding to a range resolution of 4.26 cm. In some embodiments, the CIR measured by the nth antenna in a time slot can be expressed as: (Formula 1) TIFF2026009905000002.tif25135Here, L is the number of range taps, N is the number of antennas, δ(·) is the Delta function, g n,l and τ l are the complex channel gain and propagation delay of the l-th range tap, respectively.

[0248] The transmitted pulse signal is reflected by surrounding objects, including the fingers 191, the hand 190, the surface 101, and any other objects on the surface 101, and is finally received as a CIR by the Rx 120. In this embodiment, the virtual keyboard 130 is printed on the surface 101, which may be a flat piece of paper or a flat board. In another example, the virtual keyboard 130 may be light projected onto the surface 101. In some embodiments, the surface 101 does not need to show the entire virtual keyboard 130. As long as a user knows where to place their fingers on the surface 101 in the first place, they can type keys using the mmKey system, which can be displayed on the monitor of a computer, telephone, piano, or clock. For example, the surface 101 may display two marks representing the locations of two reference keys (e.g., F and J) on a virtual computer keyboard, and the Tx 110 and Rx 120 are positioned accordingly so that the user can type keys using the mmKey system by first finding the reference keys F and J and then typing the keys while looking at the keys displayed on the computer monitor.

[0249] For each time slot t, the captured CIR is an N × L complex matrix. By analyzing the received signals, surrounding activity can be monitored, including keystrokes by finger 191 and / or hand movements by hand 190. A coordinate system such as that shown in FIG. 1B may be used, where θ, φ, and γ denote elevation angle, azimuth angle, and range, respectively. Thus, reflected signals strike the receiving antenna array at various azimuth angles φ and elevation angles θ.

[0250] mmKey Overview. A key challenge of mmKey is quickly and robustly recognizing keystrokes from RF signals reflected not only from fingers but also from arms and other static objects. As shown in FIG. 2, the mmKey system can address this challenge by following the procedure 200. First, the system collects CIRs from the reflected signals received by the Rx in operation 210. Next, the system performs motion detection in operation 220 to adaptively and robustly detect the presence of motion. In operation 230, the system performs motion discrimination to distinguish finger keystrokes from uninteresting movements caused by hands, arms, and other objects. If a keystroke is not detected, the process returns to operation 220 to detect further motion. If the detected motion is determined to be a keystroke in operation 230, the system can optionally perform adaptive background cancellation in operation 240 to extract dynamic reflections from the mixture of superimposed reflected signals. Next, in operation 250, the system can perform keystroke localization to localize the keystroke with high resolution. In some embodiments, a single calibration can be used to map key positions at initial setup, e.g., before or during CIR collection, in operation 260, minimizing effort by requiring only three key presses. In operation 270, keystrokes are recognized based on the keystroke locations and the key position mapping generated during calibration.

[0251] Keystroke Detection and Classification -Motion Detection

[0252] In some embodiments, the initial CIR phase and amplitude are synchronized over time, so that the CIR measured by the nth antenna for the lth range tap at time t can be modeled as: (Formula 2) TIFF2026009905000003.tif12110 where, TIFF2026009905000004.tif1225 represents the change in CIR introduced by reflections from a moving target compared to the CIR at time t-1, ε n,l (t) represents the change in CIR due to measurement noise. For example, the CIR amplitude can be modeled as follows: (Formula 3) TIFF2026009905000005.tif12125 where, TIFF2026009905000006.tif1225 Reflecting contributions from TIFF2026009905000007.tif1225, TIFF2026009905000008.tif1232 is ε n,l Due to (t). TIFF2026009905000009.tif1225 It may not be exactly equal to TIFF2026009905000010.tif1229, TIFF2026009905000011.tif1229 to complex value h n,l Project in the direction of (t), Note that the same is true for TIFF2026009905000012.tif1232. Therefore, the CIR amplitude difference can be calculated as follows: (Formula 4) TIFF2026009905000013.tif12158

[0253] In the absence of motion, i.e., in Eq. TIFF2026009905000014.tif1237, and in Eq. TIFF2026009905000015.tif1237, based on Eq. 4 TIFF2026009905000016.tif1266. Without loss of generality, measurement noise We can assume that the amplitude change caused by only TIFF2026009905000017.tif1232 follows a Gaussian distribution. Next, when there is no motion, Δ|hn,l By collecting a sequence of |(t)|, Δ|h ref,n,l We can construct a "still" frame, denoted as |h|, and use the Z-score anomaly detection method to find the CIR amplitude difference Δ|h of the input n,l (t)| and Δ|h ref,n,l By comparing |, movement can be detected in real time.

[0254] More specifically, the sample mean and sample standard deviation Δ|h ref,n,l |Δ|h n,l (t)| by centering and normalizing Δ|h n,l The Z-score of (t)| can be evaluated as follows: (Formula 5) TIFF2026009905000018.tif2088 where, TIFF2026009905000019.tif1447 is Δ|h ref,n,l are the sample mean and standard deviation of |Z n,l The larger the value of (t), the more likely the sample is to diverge from the reference frame, and the more likely motion occurs at time t.

[0255] Figure 3A shows the reference frame Δ|h ref,n,l An example of | is shown below. Z-score based anomaly detection can assume that the reference sample sequence follows a Gaussian distribution. Therefore, as shown in Figure 3B, Δ|h ref,n,l We can examine the quantile-quantile (QQ) plot of the normalized sample of |h ref,n,l | is very close to a normal distribution and satisfies the requirements for Z-score calculation. Figure 3C shows the Δ|h n,l (t)|. Each time there is a keystroke, Δ|h n,l (t)| causes fluctuations, which are shown by the dotted line in Figure 3C, at the threshold calculated by Eq. (5). n,l This can be captured by evaluating (t)|

[0256] Motion detection aims to detect the start and end times of keystrokes and their corresponding ranges. Instead of relying on Z-scores calculated from one single antenna, all available antennas and range taps can be utilized to improve robustness. In some embodiments, a sliding window with length W is applied to the input CIR stream, and for each window, the CIR can be obtained as a complex-valued matrix N×L×W, where N and L are the number of antennas and range taps, respectively. The corresponding value |Δh n,l By adopting a majority vote, we can construct an indicator matrix I(t) with dimensions N×L×W, where each element I n,l (t)=1{Z n,l (t)>υ}, where 1 is the indicator function, υ=3 is a value commonly used for Z-score anomaly detection, and |Δh n,l This means that a difference of more than 3 standard deviations of |Z| can be detected. n,l Comparing (t) and υ means that Δ|h n,l (t)|. Then, if the majority of the elements of I(t) are 1, then motion is detected for the current window. Then, the motion range taps are The start or end point of the movement can be further estimated as satisfying TIFF2026009905000020.tif20101. This can be determined by searching for the first and last anomalous time slots on the taps of TIFF2026009905000021.tif1428.

[0257] Keystroke Identification

[0258] Although a motion detector can identify which range taps are affected by a motion, it cannot distinguish whether the motion is caused by a keystroke or a hand movement. Distinguishing between keystrokes and hand movements can arise from two observations: 1) hand movements are usually accompanied by some shift in hand position, whereas finger keystrokes are not, and 2) hand movements can affect a much larger reflection area than finger movements. Therefore, two features can be devised to distinguish between keystrokes and hand movements: CIR phase and dynamic level.

[0259] Raw CIR Phase: Compared to CIR amplitude, CIR phase is more sensitive to small position shifts of the reflector. In some embodiments, CIR phase is already synchronized between all antennas and all samples. For example, for a carrier frequency operating at 60.48 GHz, the wavelength is λ = c / f = 5 mm, which means that a small radial shift of the reflector of 2.5 mm toward / away from the radio will cause a 2π change in CIR phase, supporting accurate classification of gross (e.g., hand) and fine (e.g., fingertip) movements.

[0260] Figure 4A shows the amplitude difference of the CIR, Δ|h n,l (t)|, and Figure 4B shows the CIR phase ∠h from a CIR sequence containing three palm movements indicated by the darker rectangles. n,l (t), each followed by a single finger keystroke indicated by a lighter line rectangle. n,l All six movements can be detected based on Δ|h n,l From (t)|, it is difficult to know whether the movement is a finger keystroke, which is ∠h n,lIt may be more discriminative by measuring (t). As shown in Figure 4B, hand movements produce much higher peaks due to larger position changes than finger keystrokes. Therefore, peak height acts as a promising feature for distinguishing these two movements. As shown in Figure 4C, which shows the peak height of the CIR raw phase, we can define the peak height as the average height of the heights on either side of the peak. Because hand shifts impact more antennas and may cross multiple taps, we can calculate the CIR phase ∠h across all antennas. n,l (t) and three adjacent taps centered on the target tap, i.e. TIFF2026009905000022.tif1146 (corresponding to an area of ​​approximately 13 cm, for example) can be merged.

[0261] Dynamic Level: By observing that hand movements affect a larger reflex area than finger keystrokes, we can uncover a novel feature, dynamic level, to describe such differences. Dynamic level may be defined as the ratio of non-DC power to total power in the CIR. In terms of γ, it can be written as: (Formula 6) TIFF2026009905000023.tif2595 where H l,n (f)=FFT(h l,n (t)). The denominator is the total power of the signal reflected from both the static background and the dynamic hand / fingers. The numerator is the power reflected only by moving objects (excluding the DC component). Therefore, the dynamic level increases as the size of the reflection area increases. In other words, hand movements can result in higher dynamic levels than finger movements. Figure 4D shows the distribution of dynamic levels for one-finger keystrokes, two-finger keystrokes, three-finger keystrokes, and hand movements, respectively. As shown, the three types of keystrokes share similar dynamic levels, while hand movements experience much larger values, providing it as an effective metric for distinguishing between hand and finger movements.

[0262] Combining these two features together allows for motion discrimination with a simple two-step verification. For example, once motion is detected and segmented, CIR frames are evaluated by thresholding the peak height of the raw CIR phase and then thresholding the dynamic level. In some embodiments, only motions with both low peak height and low dynamic level are considered finger keystrokes. Experiments have shown that this conservative decision rule can completely filter out hand interference motions with an empirical preset threshold, but it may also cause some misdetection of finger keystrokes, which is measured by detection accuracy and later evaluated. In other embodiments, motions with low peak height or low dynamic level can be considered finger keystrokes.

[0263] Keystroke location

[0264] Adaptive background cancellation: As shown in Figure 5A, the received signal is a mixture of reflections from all dynamic and static objects. Therefore, we can remove the background reflections and extract only the dynamic component associated with the keystroke.

[0265] According to equation (2), for each time slot t, CIR h n,l (t) is CIR h n,l (t-1) and their difference. From t-1 to t, the reflection from the static background is h n,l (t-1), and the change in CIR is a new dynamic reflection. Component and noise ε from TIFF2026009905000024.tif1225 n,l (t) and the component by (t). Therefore, the term h n,l The effect of background reflections can be eliminated by subtracting (t-1). As shown in Figure 5B, for M consecutive samples, Assuming that TIFF2026009905000025.tif1225 does not undergo significant changes, TIFF2026009905000026.tif1225 can be estimated as follows: (Formula 7) TIFF2026009905000027.tif25103Here, M indicates the number of samples used for background cancellation.

[0266] Keystroke location

[0267] Location by MUSIC

[0268] After extracting the dynamic signal provided by a finger keystroke, three-dimensional (3-D) coordinates of the keystroke location can be obtained, which can be translated to the actual key, as described in more detail below. Spatial resolution can be limited by the small effective aperture of the receiving antenna array. In some embodiments, to improve spatial resolution and thus accurately localize keystrokes, the mmKey system performs digital beamforming on the received CIR based on the MUltiple Signal Classification (MUSIC) algorithm. The basic idea of ​​the MUSIC algorithm is to perform eigenvalue decomposition on the covariance matrix of the CIR, resulting in a signal subspace that is orthogonal to the noise subspace. MUSIC is used to reconstruct the spatial spectrum of sparse signals, with the goal of typically locating fewer than 10 keystrokes.

[0269] In the example below, the target estimated in the previous module We focus on the range taps of TIFF2026009905000028.tif1428. In the coordinate system shown in Figure 1B, we assume that there are D reflected signals impinging on the receiving antenna array with different azimuth angles φ and elevation angles θ. Then, the CIR h can be formulated as follows: (Formula 8) TIFF2026009905000029.tif23125Here, S(θ i ,φi ) is (θ i ,φ i ) and corresponds to the direction of the ith reflected signal, i.e., the direction (θ i ,φ i ) is the normalized phase response of the antenna array to a signal coming from x i represents the complex value of the i-th reflected signal, and ε j represents the additive thermal noise due to the jth antenna, which is assumed to be a Gaussian random variable with mean 0 and independent and identically distributed (IID) for different receive antennas. N is the number of antennas. A more concise matrix representation of equation (8) can therefore be written as h = Sx + ε, where S is defined as the steering matrix. The covariance of h can then be evaluated as, (Formula 9) TIFF2026009905000030.tif11137 where, TIFF2026009905000031.tif1144 and Rs and Rε are the covariance matrices of the signal and noise components, respectively. Then, the eigenvalue decomposition can be expressed as follows: (Formula 10) TIFF2026009905000032.tif1988 where Us is the signal space, while Uε is the noise space. The MUSIC spatial spectrum can be expressed as: (Formula 11) TIFF2026009905000033.tif13103

[0270] Figure 6 shows the pseudo-spatial spectrum of a single keystroke movement. A peak in the spatial spectrum P indicates the presence of a reflected signal from a finger keystroke, while a low value of P indicates the absence of such a reflection. Other spectral estimation methods, such as conventional beamforming (CBF) or minimum variance distortion response (MVDR) beamforming (also known as Capon beamforming), can also be employed.

[0271] Refine Position

[0272] The MUSIC algorithm can achieve high resolution when determining the position of the source of movement, but actually requires prior knowledge of the number of sources, which is not usually known. To address this issue, a peak selection module can be applied before target localization. A predetermined number of targets K are supplied to the MUSIC algorithm to obtain an initial pseudo-spectrum. K peaks will be extracted from the pseudo-spectrum regardless of the actual number of existing targets. Next, the peak selection module is designed to remove false peaks.

[0273] Figures 7A - 7C show the pseudo-spectra of the movement of a single keystroke with different numbers of targets K. As K increases, there are increasingly more outlier peaks, including (i) lower peaks in the background and (ii) higher peaks spreading from the target peaks.

[0274] To remove false peaks and determine the number of true targets, two criteria can be followed. First, peaks with heights lower than a pre-set adaptive threshold th1 are regarded as noise peaks and removed by a filter. Since it is common, th1 may be a proportional function of the height of the highest peak, i.e., th1 = c·max(p1,..., p k ) which is determined in the calibration stage. Second, for adjacent spreading peaks, when the angular spatial distance between these peaks is within the threshold d th , agglomerative hierarchical clustering is applied to merge them. Observing that due to the signal reflected by the upper part of the finger, peaks tend to expand more in the elevation direction as shown in Figures 7B and 7C, a relatively small weight can be applied in the elevation direction. More specifically, the distance between two peaks (Δθ, Δφ) is weighted by (a, b) respectively, where a < b to allow more expansion of the peak broadening at θ. dth is an adaptive threshold indicating the size of the cluster, which is a function of the distance between the wireless device and the keyboard, i.e., Behaves as a function of TIFF2026009905000034.tif1110.

[0275] After filtering and clustering the detected peaks, the number of keystrokes is estimated as the number of clusters, and the highest peak in each cluster is considered as the representative of the cluster. The estimated location, shown as TIFF2026009905000035.tif1224, is fed into a keystroke recognition module, described below.

[0276] Keystroke Recognition

[0277] Finger keystroke locations estimated by the super-resolution MUSIC algorithm TIFF2026009905000036.tif1224 can only reflect the relative position of the keystroke to Rx. Knowledge of the keyboard position relative to Rx may be needed to map the keystroke position to the keyboard and infer which key is being pressed. This allows for position TIFF2026009905000037.tif1224 keystrokes can be translated to specific keys. To obtain such a mapping relationship, a simple calibration process can be employed, which only needs to be done once during keyboard initialization. Because mmKey is compatible with various types of keyboards, including piano keyboards and computer keyboards, the following examples start with the one-dimensional (1-D) case using the white keys of a piano keyboard as an example, and later extend to the common two-dimensional (2-D) case for computer keyboards and telephone keypads.

[0278] 1-D Case. To complete keyboard calibration with minimal effort, the user can randomly select and press three keys. As shown in Figure 8A, keys w1, w6, and w10 are pressed during calibration, and the corresponding estimated azimuth angles by the MUSIC algorithm are Assuming it is represented by TIFF2026009905000038.tif1449, then TIFF2026009905000039.tif1499 can be obtained. As shown in Figure 8A, according to Sin's law, we can have the following formula: (Formula 12) TIFF2026009905000040.tif4474Here, β1 and β2 are two unknown angles that belong to two adjacent triangles and form a straight angle. From equation (12), denoting the ratio of |AC| to |BC| as η, we can have the following equation: (Formula 13) TIFF2026009905000041.tif1967

[0279] In this example, the ratio |AD| / |BD| is already known as 5 / 4 according to the keyboard layout. We can derive the value of η by assuming that all keystrokes occur at the center of the key. We can also derive the azimuth angle boundary between every two adjacent keys. For example, to calculate the boundary between keys w2 and w3, as shown in Figure 8B, we can again apply Sin's law as follows: (Formula 14) TIFF2026009905000042.tif4573 here TIFF2026009905000043.tif1454 is the angle corresponding to the angle in Figure 8B. Then, the following equation is obtained: (Formula 15) TIFF2026009905000044.tif2576

[0280] Based on equation (15), according to the keyboard layout, the ratio |AE| / |BE| is known as 1 / 5, so You can get the exact value of TIFF2026009905000045.tif1435. Similarly, you can get the exact value of (w1, w2), ..., (w9, w 10) can be derived. By subtracting the absolute azimuth of w1, the azimuth of the boundary can be calculated for keystroke recognition.

[0281] 2-D Case. The geometric model from the 1-D case can be extended to two dimensions, and both elevation and azimuth angles are used for keystroke recognition. As shown in Figure 9, three keys, "1," "G," and "M," are pressed for calibration. In the horizontal azimuth direction, we have a triangle ΔA1B1C1, from which all azimuth boundaries of the keys can be derived. Meanwhile, in the vertical elevation direction, we have a triangle ΔA2B2C2, which can be used to calculate the elevation boundaries. Here, C1 and C2 represent the same location on the device in terms of the azimuth and elevation dimensions, respectively. Given the boundary azimuth and elevation angle values ​​for each key, keystrokes can be easily recognized in real time by mapping the estimated keystroke locations to target keys on a keyboard with a known layout.

[0282] Experimental Evaluation: mmKey can be prototyped and real-world experiments can be performed using a testbed that repurposes an 802.11ad / ay chipset as a radar-like platform. The device is placed over a flat surface supporting a printed virtual keyboard. Various types of keyboards can be considered, such as a QWERTY computer keyboard, piano keys, or smartphone keypad. For each keyboard, a layout may be printed on paper, maintaining the same physical dimensions so that the user has a typing experience as familiar as a real keyboard. In some embodiments, the distance between the keyboard and the device is set to 20 cm to ensure that the keyboard is within the device's field of view (FoV), which is 100° in both azimuth and elevation. The default sampling rate is f s =1 / T b = 100 Hz, where T b is the burst duration as shown in Figure 1C.

[0283] Three main metrics can be used for evaluation. Detection accuracy (DA) and recognition accuracy (RA) are defined to quantify how well mmKey detects keystrokes and how well it localizes and recognizes them, respectively. Based on DA and RA, over all accuracy (OA) is calculated as OA = DA × RA. DA and RA are defined as follows: (Formula 16) TIFF2026009905000046.tif27109

[0284] We first investigated the performance of mmKey on a virtual computer keyboard. A standard printed alphanumeric keyboard has a typical QWERTY-based layout with a distance of 19 mm between adjacent keys. The key involved is the letter and number keys in the experiment, selecting "1," "G," and "M" as landmark keys for calibration. Once setup is complete, each participant repeatedly presses a predefined sequence of keys. Figure 10A shows the OA confusion matrix for recognizing 36 keys (26 letters + 10 numbers) on a virtual computer keyboard. In this embodiment, mmKey achieves remarkable keystroke recognition with an average OA of 95.42% on a computer keyboard. In some situations, some samples of a key, especially samples below the actual key, are recognized as adjacent keys. This is due to the presence of reflections from the knuckles, which lead to estimation errors in the elevation angle direction. In practical applications where users are typing typical text, these errors can be easily recovered by robust spell-checking techniques.

[0285] Word Recovery: Furthermore, mmKey's ability in recovering input sentences is investigated and accuracy is evaluated at the word level. In this example, by collecting sentence samples, users are asked to type each of the following sentences five times on a printed computer keyboard: S1 = "The quick brown fox jumped over the lazy dog", S2 = "No one knew why the candle went out", S3 = "Autumn leaves look like golden snow", S4 = "It's not as deep as you think", and S5 = "My little pet mouse escaped from his cage".

[0286] First, we can run mmKey on the CIR data and get the direct output, i.e., the sequence of recognized keys. Then, we can feed the output to Grammarly, a popular commercial English transcription tool, for correction. Here, we define word-level accuracy (WA) as The accuracy can be calculated using the TIFF2026009905000047.tif1985 algorithm, and the results are shown in Figure 10B. Because one misrecognized character leads to a misrecognized word, the WA on mmKey's direct output, approximately 80%, is expected to be less than its OA. With the help of spell checking / text correction, word-level misrecognitions can be easily corrected with a considerable accuracy rate of over 97%. With this high accuracy, mmKey could promise to become a ubiquitous virtual keyboard for practically anywhere, mobile, and portable use.

[0287] Virtual piano keyboard. Considering both white and black keys, Fig. 11A shows the confusion matrix for single-key keystrokes, where "?" denotes a missed detection. As shown in Fig. 11A, mmKey achieves a high OA of 99.12%.

[0288] To play the piano, a user may need to press multiple keys simultaneously. As shown in Figure 11B, when a user presses two simultaneous keys, mmKey accurately recognizes keystrokes when the two keys are located far enough apart. However, when the two pressed keys are close together, especially when they are adjacent keys, accuracy may decrease due to co-located fingers. In this example, the overall accuracy of double-key keystroke recognition is 92.54% for all cases, and the accuracy is 96.93% for non-adjacent keys. Observing accuracy along the diagonal, we can see that OA decreases near the edges of the keyboard due to the effect of inter-finger blockage at the edge positions.

[0289] Figures 12A-12B show multi-keystroke accuracy, with Figure 12A showing accuracy versus number of keystrokes and Figure 12B showing the confusion matrix for detection. For three or more simultaneously pressed or typed keys, results show that OA decreases to 76.67% when there are three keys pressed or typed, and further decreases to 65.94% for four keys. Detection accuracy is shown in Figure 12B, where more keystrokes result in more miss-detections due to interference between multiple fingers, but as illustrated in Figure 12A, this does not significantly affect recognition accuracy. Once a keystroke is detected, mmKey is able to accurately recognize it.

[0290] FIG. 13 shows a flowchart of an example method 1300 for wirelessly tracking keystrokes according to some embodiments of the present disclosure. At operation 1302, a first wireless signal is transmitted over a wireless channel at a location including a surface. At operation 1304, a second wireless signal is received over the wireless channel, where the second wireless signal includes a reflection of the first wireless signal by at least one moving object at the location. At operation 1306, a time series of channel information (CI) for the wireless channel is obtained based on the second wireless signal, where each CI includes at least one of channel state information (CSI), a channel impulse response (CIR), a channel frequency response (CFR), or a received signal strength index (RSSI). At operation 1308, at least one keystroke on the surface is detected based on the time series of CI (TSCI). At operation 1310, at least one location of the at least one keystroke on the surface is determined. At least one key associated with the at least one keystroke is determined based on the at least one location in act 1312. The order of the acts in Figure 13 may be changed according to various embodiments of the present disclosure.

[0291] In some embodiments, the wireless keystroke tracking method includes steps s1-s7 described below.

[0292] Step s1: Using one transmit (Tx) antenna and multiple receive (Rx) antennas (e.g., h(Rx antenna, distance)), a time series of CIR is captured.

[0293] Step s2: Detect the presence of motion, including steps s2a to s2d. Steps s2a and s2b are executed for each receiving antenna.

[0294] Step s2a: Calculate the differential magnitude of the empty CIR based on the frame of the CIR in the empty case (e.g., empty CIR Δ|h(Rx antenna, distance)|). For each Rx antenna and distance, calculate its sample mean and standard deviation (u and s).

[0295] Step s2b: At each time instance, calculate the differential magnitude of the target CIR (e.g., Δ|h(Rx antenna, distance)| of the target CIR). For each receive antenna and distance, calculate a z-score. Compare the z-score with a threshold T1. If the z-score at time t is greater than Ta, motion is detected for that antenna and distance. A majority vote is then taken across all Rx antennas. The z-score is calculated by taking the difference between Δ|h(Rx antenna, distance)| and u, and normalized by s as in Equation 5. The threshold T1 is an empirical threshold, and 3 is a commonly used number for T1.

[0296] Step s2c: Determine the start and end times of the movement by identifying the first and last time instances when the majority of the Rx antennas detect movement.

[0297] Step s2d: Determine the maximum movement distance (sum of z-scores over all antennas and movement durations).

[0298] In some embodiments, motion can be detected by thresholding the variance of the CIR amplitude |h(Rx antenna, distance)|, using the assumption that the variance of noise is much smaller than the variance of motion. The target range where motion occurs can then be determined by maximizing the variance over the range dimension.

[0299] Step s3: Finger movement identification, including steps s3a and s3b.

[0300] In step s3a: For each target distance and distances adjacent to the target distance (e.g., l-1, l, l+1), steps s3a1 and s3a2 ​​are performed. Step s3a1: Calculate the peak height of the raw CIR phase and compare it to a threshold T2. Step s3a2: Calculate the dynamic level and compare it to a threshold T3. The thresholds T2 and T3 are determined during the calibration phase, and the dynamic level is calculated by equation (6).

[0301] In step s3b: Perform majority voting for all antennas and distances. If the majority of peak heights are less than T2 and the dynamic level is less than T3, the movement is recognized as a keystroke.

[0302] In some embodiments, the duration of a movement can be used as a feature to distinguish a keystroke from other movements, given that a keystroke is a very rapid movement of short duration. In some embodiments, the sum of the variance values ​​of all antennas can be used to distinguish a keystroke from other movements, given that hand movements affect more antennas than finger movements.

[0303] Step s4: For the CIR associated with each keystroke movement at the target distance l, extract the time-varying component of the CIR and perform background subtraction (e.g., Δh), including steps s4a and / or s4b. Step s4a: Background subtraction may be performed by successive CIR subtraction. Step s4b: Background subtraction may be performed by subtracting the time average of several preceding CIR frames.

[0304] Step s5: Run MUSIC on Δh using a preset number of keystrokes K to obtain a 2D pseudospectrum P(theta, phi). The pseudospectrum is refined by thresholding out the lower spectral peaks and collecting the diffuse peaks. The final angular direction (theta * , Phi *) is obtained. Step s5 includes steps s5a to s5c. Step s5a: K is determined empirically. In the experiment, K = 5. Step s5b: Peaks with heights lower than T4 are considered as noise peaks and are filtered out. T4 is determined in the calibration stage. In step s5c, peaks are considered as diffuse peaks and are filtered out if the distance between the peak and the highest peak is less than a threshold T5. The distance is calculated in 2-D space with more weight placed on the elevation dimension (e.g., theta). T5 is determined in the calibration stage. The weights in (theta, phi) are (2, 1) in the experiment.

[0305] In some embodiments, beamforming can be performed using conventional beamforming (CBF) and / or minimum variance distortion response (MVDR). In some embodiments, adaptive subspacing in MUSIC or other heuristic peak selection methods can be used to derive the number of sources to refine the pseudospectral peaks.

[0306] Step s6: Calibration is performed, including one or more of steps s6a-s6h. In some embodiments, during initial setup, the keyboard is calibrated by pressing three different keys. The angular boundaries of adjacent keys are calculated based on the geometric relationships shown in Equations 13 to 15 (assuming r is the same based on the Tx / Rx placement relative to the keyboard). Step s6a: For keyboards with a 2-D placement, the three keys are placed on different rows and columns (e.g., (theta, phi)). Step s6b: The calibration phase includes both finger movements (keystrokes) and hand movements (movements from one keystroke to another). Step s6c: The target distance and number of keystrokes are known during the calibration phase. Step s6d: T2 may be calculated by averaging the CIR phase peak heights of the keystroke movements and hand movements at the target distance. Step s6e: T3 may be calculated by averaging the dynamic levels of the keystroke movements and hand movements at the target distance. In step s6f, T4 may be calculated by averaging the noise peak height and keystroke peak height in the pseudospectrum. In step s6g, T5 may be calculated by the maximum distance of the diffusion peak to the highest peak. In step s6h, T1 through T5 may be manually determined by observing the relevant parameters within the calibration phase.

[0307] In step s7, the keystroke (theta) in step s5 is * , Phi * ) to the keyboard according to the angle bounds in step s6.

[0308] The following numbered sections provide examples for wirelessly tracking keystrokes.

[0309] Clause 1. A system for wirelessly tracking keystrokes on a surface, comprising: a transmitter configured to transmit a first wireless signal over a wireless channel of a location including the surface using a transmit antenna; a receiver configured to receive a second wireless signal over the wireless channel using a plurality of receive antennas, the second wireless signal comprising a reflection of the first wireless signal by at least one moving object in the location; and a processor configured to obtain, for each of the plurality of receive antennas, a time series of channel information (CI) for the wireless channel based on the second wireless signal, each CI representing a channel and a processor configured to: acquire, based on a time series of CIs (TSCIs) obtained for each of the plurality of receive antennas, at least one of channel state information (CSI), a channel impulse response (CIR), a channel frequency response (CFR), or a received signal strength index (RSSI); detect at least one keystroke on the surface based on the time series of CIs (TSCIs) obtained for each of the plurality of receive antennas; determine at least one location of the at least one keystroke on the surface; and determine at least one key associated with the at least one keystroke based on the at least one location.

[0310] Clause 2. The system of clause 1, wherein each CI includes a CIR, the first radio signal is carried by millimeter waves, and each of the at least one moving object is a finger or tip configured to type and perform keystrokes on a virtual keyboard, and has a position determined based on a plurality of spatial bins of the location, each of the plurality of spatial bins being determined by a respective angular direction and a respective distance range emanating from the receiver, and each angular direction being identified by a corresponding azimuth angle and a corresponding elevation angle.

[0311] Clause 3. The system of clause 2, wherein detecting the at least one keystroke includes detecting a movement of each of the at least one moving object on the surface based on a time series of the CIs (TSCIs) acquired for each of the plurality of receiving antennas, and recognizing the movement as a keystroke performed by the moving object.

[0312] Section 4. The system described in Section 3, wherein detecting the movement of each moving object includes, for each time instance, calculating, for each of a plurality of distance ranges of interest and each of the plurality of receiving antennas, a differential CIR for the distance range of interest based on a CIR amplitude measured by the receiving antenna at the time instance and a CIR amplitude measured by the receiving antenna at a preceding time instance before the time instance; calculating, for the receiving antenna and the distance range of interest, a measurement score based on the differential CIR, a sample mean of reference differential CIRs, and a standard deviation of the reference differential CIRs; comparing the measurement score with a first threshold; and determining that candidate movement has been detected by the receiving antenna in the distance range of interest if the measurement score is greater than the first threshold.

[0313] Clause 5. The system of clause 4, wherein the reference differential CIR is calculated based on the amplitude of a reference CIR measured by the receiving antenna at two consecutive time instances, the reference CIR being the CIR of the wireless channel, and the reference CIR being obtained without any moving objects at the location.

[0314] Clause 6. The system of clause 4, wherein detecting the motion of each moving object further includes, for each candidate motion, determining that the candidate motion is target motion if the candidate motion is detected by a majority of the plurality of receiving antennas at a time instance.

[0315] Clause 7. The system of clause 6, wherein detecting the motion of each moving object further includes: for each target motion, determining a start time of the target motion based on a first time instance at which a majority of the plurality of receive antennas detects the target motion; and determining an end time of the target motion based on a last time instance at which a majority of the plurality of receive antennas detects the target motion.

[0316] Clause 8. The system of clause 7, wherein detecting the movement of each moving object further includes determining, for each target movement, a target distance range for the target movement based on a distance range that maximizes a sum of measurement scores calculated for all of the plurality of receiving systems and all time instances from the start time to the end time.

[0317] Item 9. The system of item 8, wherein recognizing the movements as keystrokes includes, for each target movement, for each of the plurality of receive antennas, and for each of three spatially consecutive distance ranges centered on the target distance range of the target movement, calculating a peak height of the CIR phase measured by the receive antenna for the distance range, calculating a first total power of non-zero frequency components of the CIR signal measured by the receive antenna for the distance range, and calculating a second total power of all frequency components of the CIR signal measured by the receive antenna for the distance range.

[0318] Item 10. The system of item 9, wherein recognizing the movement as a keystroke further includes: for each target movement, calculating an aggregate peak height based on peak heights of CIR phases measured by all of the multiple receiving antennas for all of the three spatially consecutive distance ranges; comparing the aggregate peak height with a second threshold; calculating a first aggregate power based on a first total power calculated for all of the multiple receiving antennas and all of the three spatially consecutive distance ranges; calculating a second aggregate power based on a second total power calculated for all of the multiple receiving antennas and all of the three spatially consecutive distance ranges; calculating a dynamic level representing a reflective area associated with the target movement based on a ratio between the first aggregate power and the second aggregate power; comparing the dynamic level with a third threshold; and recognizing the target movement as a keystroke movement if the aggregate peak height is smaller than the second threshold and the dynamic level is smaller than the third threshold.

[0319] Clause 11. The system of clause 10, wherein each of the second and third thresholds is predetermined based on a calibration of the system, the calibration relating to keystrokes each corresponding to a known target distance range and hand movements each corresponding to moving from one keystroke to another, the second threshold being calculated by averaging peak heights of CIR phases of keystrokes and hand movements in the target distance range, and the third threshold being calculated by averaging dynamic levels of keystrokes and hand movements in the target distance range.

[0320] Item 12. The system of item 9, wherein recognizing the movement as a keystroke further includes: for each target movement and for each of a plurality of antenna-distance groups, each antenna-distance group including a respective one of the plurality of receiving antennas and a respective one of three spatially contiguous distance ranges centered on the target distance range of the target movement, comparing the peak height with a second threshold; calculating a dynamic level representing a reflective area associated with the target movement based on a ratio between the first total power and the second total power; and comparing the dynamic level with a third threshold, each of the second threshold and the third threshold being predetermined based on calibration of the system; recognizing the target movement as a candidate keystroke movement if the peak height is smaller than the second threshold and the dynamic level is smaller than the third threshold; and recognizing the target movement as a keystroke movement if, for each target movement, the target movement is recognized as a candidate keystroke movement by a majority of the plurality of antenna-distance groups.

[0321] Clause 13. The system of clause 12, wherein the processor is further configured to calculate, for each CIR, at each time instance, in association with each keystroke movement, a background-subtracted CIR by at least one of subtracting from the CIR the CIR of a preceding time instance before the time instance, or subtracting from the CIR a time average value of multiple CIRs of multiple preceding time instances before the time instance.

[0322] Clause 14. The system of clause 13, wherein determining the at least one location of the at least one keystroke includes: for each background-subtracted CIR associated with each keystroke movement in a target distance range, applying digital beamforming to the background-subtracted CIR based on MUSIC (MUltiple SIgnal Classification) and a predetermined number K representing a maximum amount of keystrokes that can occur simultaneously at the location; calculating a two-dimensional spatial spectrum based on the digital beamforming, the spectrum including CIR power as a function of different positions, each represented by an angular direction including an azimuth angle and an elevation angle in the target distance range; and identifying K peaks in the two-dimensional spatial spectrum corresponding to the K highest CIR powers.

[0323] Clause 15. The system of clause 14, wherein determining the at least one position of the at least one keystroke further includes: refining the K peaks by at least one of removing from the K peaks peaks whose heights are lower than a fourth threshold or removing from the K peaks diffuse peaks whose distances to a highest peak of the K peaks in the two-dimensional spatial spectrum are less than a fifth threshold to obtain at least one refined peak; and determining, based on each of the at least one refined peak, one corresponding to the at least one position of the at least one keystroke.

[0324] Item 16. The system of item 15, wherein each of the fourth threshold and the fifth threshold is predetermined based on calibration of the system, the fourth threshold is calculated by averaging noise peak heights and keystroke peak heights in the two-dimensional spatial spectrum, and the fifth threshold is calculated based on the maximum distance from all diffusion peaks to the highest peak in the two-dimensional spatial spectrum.

[0325] Clause 17. The system of clause 15, wherein the transmitter is further configured to transmit three wireless signals through the wireless channel at the location during calibration of the system using the transmitting antenna, and the receiver is further configured to receive three reflected wireless signals through the wireless channel using the multiple receiving antennas, wherein each of the three reflected wireless signals includes a reflection of a corresponding one of the three wireless signals when a corresponding one of three keystrokes corresponding to three known keys is performed on the virtual keyboard, and the processor is further configured to obtain, for each of the multiple receiving antennas, a TSCI of the wireless channel based on each of the three reflected wireless signals, determine three positions on the virtual keyboard that correspond to each of the three known keys based on the corresponding TSCI, and calculate angular boundaries of adjacent keys on the virtual keyboard based on the three positions of the three known keys and geometric relationships between different keys on the virtual keyboard.

[0326] Clause 18. The system of clause 17, wherein the virtual keyboard has a two-dimensional layout, and the three known keys are located in different rows and different columns of the virtual keyboard.

[0327] Item 19. The system of item 17, wherein determining the at least one key associated with the at least one keystroke includes, for each of the at least one position, mapping an angular orientation representing the position to the virtual keyboard according to the angle bounds calculated during the calibration to determine a corresponding key on the virtual keyboard.

[0328] Clause 20. The system of clause 14, wherein determining the at least one location of the at least one keystroke further includes refining the K peaks based on an adaptive subspacing method or a heuristic peak selection method to obtain at least one refined peak, and determining a corresponding one of the at least one location of the at least one keystroke based on each of the at least one refined peak.

[0329] Item 21. The system of item 13, wherein determining the at least one location of the at least one keystroke includes: for each of the background-subtracted CIRs associated with each keystroke movement in a target distance range, applying beamforming to the background-subtracted CIRs based on a minimum variance distortion response (MVDR); calculating a two-dimensional spatial spectrum including CIR power as a function of different positions, each represented by an angular direction including an azimuth angle and an elevation angle in the target distance range, based on the beamforming; and identifying K peaks in the two-dimensional spatial spectrum corresponding to the K highest CIR powers, where K is a predetermined amount of keystrokes that can occur simultaneously at a location.

[0330] Item 22. The system of item 7, wherein recognizing the movements as keystrokes includes, for each target movement, determining a time duration between a start time of the target movement and an end time of the target movement, and recognizing the target movement as a keystroke if the time duration is less than a predetermined threshold.

[0331] Clause 23. The system of clause 3, wherein detecting the motion of each moving object includes determining that target motion is detected by a majority of the plurality of receiving antennas based on a comparison between a variance of CIR amplitudes measured by each of the plurality of receiving antennas and a predetermined threshold, and determining a target distance range for the target motion based on a distance range that maximizes the variance of the CIR amplitudes.

[0332] Clause 24. The system of clause 23, wherein recognizing the movement as a keystroke includes: calculating, for each target movement, a sum of the variances of CIR amplitudes measured by all of the plurality of receive antennas; and recognizing the target movement as a keystroke movement if the sum is less than a predetermined threshold.

[0333] Clause 25. The system of clause 2, further comprising a display of the virtual keyboard on the surface, wherein the processor is further configured to determine a user's intended input to at least one of a computer, a piano, a telephone, or a clock based on the at least one key.

[0334] Clause 26. The system of clause 1, wherein the transmitter and the receiver are physically coupled to each other on the same side of the surface.

[0335] Clause 27. A wireless device of a wireless tracking system, comprising: a processor; a memory communicatively coupled to the processor; and a receiver communicatively coupled to the processor, wherein the additional wireless device of the wireless tracking system is configured to transmit a first wireless signal through a wireless channel at a location including a surface, the receiver being configured to receive a second wireless signal via the wireless channel, the second wireless signal including a reflection of the first wireless signal by at least one moving object at the location, and the processor is configured to: acquire a time series of channel information (CI) for the wireless channel based on the second wireless signal, each CI including at least one of channel state information (CSI), channel impulse response (CIR), channel frequency response (CFR), or received signal strength index (RSSI); detect at least one keystroke on the surface based on the time series of CI (TSCI); locate at least one location of the at least one keystroke on the surface; and determine at least one key associated with the at least one keystroke based on the at least one location.

[0336] Clause 28. A wireless device as described in clause 27, wherein each CI includes a CIR, and each of the at least one moving object is a finger or tip configured to type on a virtual keyboard to perform keystrokes, and has a position determined based on a plurality of spatial bins within a location, each of the plurality of spatial bins being determined by a respective angular direction and a respective distance range emanating from the wireless device, and each angular direction being identified by a corresponding azimuth angle and a corresponding elevation angle.

[0337] Clause 29. The wireless device of clause 28, wherein the additional wireless device is further configured to transmit three wireless signals through the wireless channel at the location during calibration of the wireless tracking system, and the wireless device is further configured to receive three reflected wireless signals through the wireless channel, each of the three reflected wireless signals including a reflection of a corresponding one of the three wireless signals when a corresponding one of three keystrokes corresponding to three known keys on the virtual keyboard is executed, and the processor is further configured to: obtain a TSCI of the wireless channel based on each of the three reflected wireless signals; determine three positions on the virtual keyboard, each corresponding to one of three known keys based on the corresponding TSCI; calculate angular boundaries of adjacent keys on the virtual keyboard based on the three positions of the three known keys and geometric relationships of various keys on the virtual keyboard; and for each of the at least one position, map an angular direction representing the position to the virtual keyboard according to the angle boundaries to determine a corresponding key on the virtual keyboard.

[0338] Clause 30. A method for a wireless tracking system, comprising: transmitting a first wireless signal through a wireless channel at a location including a surface; receiving a second wireless signal through the wireless channel, the second wireless signal including a reflection of the first wireless signal by at least one moving object at the location; acquiring a time series of channel information (CI) for the wireless channel based on the second wireless signal, each CI including at least one of channel state information (CSI), channel impulse response (CIR), channel frequency response (CFR), or received signal strength index (RSSI); detecting at least one keystroke on the surface based on the time series of CI (TSCI); determining at least one position of the at least one keystroke on the surface; and determining at least one key associated with the at least one keystroke based on the at least one position.

[0339] The presence of a subject affects RF signal propagation. For example, RF signals reflected from a human body are modulated by body movements, such as chest movements caused by breathing and heartbeat, allowing the subject's vital information to be revealed by analyzing the channel propagation characteristics. RF signals can be used to estimate respiration rate (RR) and heart rate (HR), but heart rate variability (HRV) cannot be obtained from RR and HR without precise timing of each heartbeat. Accurate HRV estimation is much more difficult than HR estimation. HR estimation systems typically take multiple samples in the time domain to achieve higher HR estimation accuracy, which is equivalent to averaging heartbeats over a time window. However, these require the exact timing of each heartbeat and are not applicable to HRV estimation, which involves the following challenges: First, RF signals reflected by the human chest are modulated by both breathing and heartbeat, and the distance change caused by breathing is greater than that caused by heartbeat. In signal processing terms, the signal-to-interference-and-noise ratio (SINR) is too low to recover and separate the heartbeat wave from a mixed signal. Second, the cardiac pumping motion must first pass through bones and tissue to reach the chest wall and then be detected by the RF signal. As a result, the bones and tissues of the human body act as filters, thus attenuating the signal. Therefore, the heartbeat wave captured by the RF signal lacks the sharp peaks of an electrocardiogram (ECG) signal, making it difficult to identify IBI. Furthermore, to provide a robust system for HRV estimation, the number of targets and their locations must be determined before estimating HRV for each subject, which is also nontrivial.

[0340] This disclosure describes a multi-person HRV estimation system (hereinafter referred to as "mmHRV") that uses commercial-off-the-shelf (COTS) millimeter-wave (mmWave) radios. In some embodiments, a target detector is devised to identify the number of users and their locations without prior calibration. Due to the fast attenuation of mmWave RF signals, signal strength decreases over longer distances. To detect subjects at various distances, mmHRV can utilize a two-dimensional constant false alarm detector in the range-azimuth plane to estimate the noise level, thus providing an adaptive threshold for target detection. Phase information is further used to filter out static objects (e.g., walls, furniture). There may be more than one reflection point for a single subject. As a result, to determine the number of targets, mmHRV can further employ nonparametric clustering to identify range-azimuth bins corresponding to each subject.

[0341] In some embodiments, after target detection, to estimate heart rate, it is necessary to extract a heart wave from the composite received signal, which includes the entire chest motion, including both respiratory and cardiac motion. In some embodiments, respiratory motion ranges from 4 to 12 mm at frequencies between 6 and 30 breaths per minute (BPM), while cardiac motion ranges from 0.2 to 0.5 mm at frequencies between 50 and 120 BPM, making both quasi-periodic signals. Taking advantage of this property, mmHRV utilizes a heart wave extractor that optimizes the decomposition of the composite signal into several band-limited signal components. Among the decomposed signal components, the heart wave is a wave with an amplitude and frequency that meets the requirements of a typical cardiac signal. Compared to approaches that involve sequential decomposition of the composite signal, mmHRV can avoid error propagation problems by simultaneously decomposing the signal components. Furthermore, mmHRV systems can operate in multi-user scenarios with target detection, simultaneously monitoring the HRV and / or other statistics of multiple people's heart signals.

[0342] The peaks of the estimated heart rate wave are then recognized to identify the exact time of each heart rate. As a result, the IBIs can be further derived and used to calculate commonly used HRV metrics, such as the root mean square of consecutive differences (RMSSD), the standard deviation of all IBIs (SDRR), and the percentage of consecutive IBIs that differ by more than 50 ms (pNN50).

[0343] FIG. 14A illustrates an exemplary setup for a wireless vital signs monitoring system, e.g., an mmHRV system, according to some embodiments of the present disclosure. As shown in FIG. 14A, the mmHRV system includes a device 1400, a transmitter (Tx) antenna array 1401, and a receiver (Rx) antenna array 1402. In some embodiments, the device 1400 operates in a high frequency band, such as 28 GHz, 60 GHz, or 77 GHz, with a bandwidth of 3-5 GHz. To obtain channel information, the Tx 1401 can transmit using one or more antennas, a radio signal, which is received by another Rx antenna 1402 after being reflected by objects and people at the locations illustrated in FIG. 14A.

[0344] As shown in FIG. 14A , the location where device 1400 is located may include multiple objects, including static object 1408 and subjects 1405, 1406, and 1407. Based on channel information obtained from the reflected signal by Rx 1402, the mmHRV system can simultaneously monitor the HRV of multiple people 1405, 1406, and 1407 in the location, regardless of the presence or absence of other static objects 1408 in the location. Different objects can be positioned in different directions from device 1400 without affecting the effective operation of the mmHRV system. For example, subject 1407, subject 1405, and chair 1408 are positioned at different azimuth angles from device 1400. Different objects can also be positioned at different distances from device 1400 without affecting the effective operation of the mmHRV system. For example, subject 1405 and subject 1406 are positioned at different distance ranges (but the same azimuth angle) from device 1400. Different subjects can face different directions within the location without affecting the effective operation of the mmHRV system. For example, as shown in FIG. 14A, subjects 1405, 1406, and 1407 are facing in different directions.

[0345] In some embodiments, Tx 1401 is a bot as described above, and Rx 1402 is an origin as described above. In Figure 14A, Tx 1401 and Rx 1402 are physically coupled to each other, but in other embodiments, they may be separated into separate devices. In some embodiments, device 1400 functions like a radar.

[0346] In some embodiments, to evaluate the performance of the mmHRV system, 11 participants aged 20 to 60 were asked to conduct a large-scale experiment in different settings, including different distances, orientations, and angles of incidence. Outside-of-line-of-sight (NLOS) and multi-user scenarios were also investigated. In some embodiments, experimental results showed that mmHRV achieved accurate IBI estimation with a median error of approximately 28 ms (for 96.16% accuracy). The root mean square error (RMSE) for the NLOS and multi-user cases was still within 32 ms and 69 ms, respectively. HRV metrics were also evaluated, demonstrating better performance compared to current studies. When the user was 1 meter away from the device, mmHRV achieved a mean IBI error of 3.89 ms, a mean RMSSD error of 6.43 ms, a mean SDRR error of 6.44 ms, and a mean pNN50 error of 2.52%.

[0347] The mmHRV system is a wireless system that can accurately detect a subject's heart rate signal and estimate their HRV purely by using RF signals reflected from the user's body. Figure 14B shows an exemplary processing workflow of the mmHRV system, according to some embodiments of the present disclosure. In some embodiments, the mmHRV system utilizes frequency-modulated continuous wave (FMCW) radar to transmit RF signals and capture reflections from the subject and static objects.

[0348] As shown in FIG. 14B, channel information is obtained in operation 1410 based on the reflections captured by the Rx. To detect subjects at different locations, beamforming is performed in operation 1420, for example, by a Bartlett beamformer, to obtain channel information at different azimuth range bins. Target detection is then performed in operation 1430, for example, by a target detector that adaptively estimates the noise level at various distances and azimuth angles, thereby detecting the presence of reflecting objects. The phase variance value is further utilized to distinguish subjects from static objects. A non-parametric clustering algorithm can be used to identify the number of targets and their locations.

[0349] In order to extract the heart rate signal from the phase information modulated by both respiration and heart rate, mmHRV can devise a heart rate signal extractor that can simultaneously decompose the phase signal into several narrowband signals and provide an estimate of the heart rate wave in operation 1440. The detected subject's HRV can be further analyzed in operation 1450 based on the interbeat interval (IBI) derived from the estimated heart rate signal.

[0350] Signal Model: In some embodiments, a chirp signal is transmitted by the FMCW radar, where the instantaneous transmitted frequency is a periodic linearly increasing signal, as shown in FIG. 15, and can be expressed as (Formula A1) TIFF2026009905000048.tif1944where, f c is the chirp start frequency, and T c is the chirp duration, and B is the bandwidth. With frequency modulation (FM), the transmitted signal x T (t) can be expressed as: (Formula A2) TIFF2026009905000049.tif37108Here, A T is the transmitted power. When an electromagnetic wave (EM) is reflected at a distance d(t) from the human chest, the reflected signal x R(t) can be expressed as: (Formula A3) TIFF2026009905000050.tif19147Here, A R denotes the amplitude of the received signal. d represents the round trip delay, and t d = 2d(t) / c, where c is the speed of light.

[0351] By mixing the received signal with a replica of the transmitted signal and following a low-pass filter, the channel information h(t) can be expressed as: (Formula A4) TIFF2026009905000051.tif19130 here Note that the TIFF2026009905000052.tif1925 term can be neglected, especially in short-distance scenarios. Therefore, h(t) can be written as, (Formula A5) TIFF2026009905000053.tif15108This is the frequency TIFF2026009905000054.tif1555 is a sinusoidal signal that depends on the target's range. For each chirp, the baseband signal h(t) is digitized by an analog-to-digital converter (ADC) to generate N samples per chip, called fast time. The time corresponding to the transmission of the chirp is called slow time, as shown in Figure 15. Therefore, the digitized channel information for the nth ADC sample and mth chirp can be expressed as (Formula A6) TIFF2026009905000055.tif19140Here, T f and T s are the time intervals in fast and slow time, respectively. c denotes the wavelength of the chirp.

[0352] In some embodiments of mmHRV, multiple antennas of the chipset can be utilized, using two Tx and four Rx antennas, as shown in Figure 16. To increase the azimuth resolution, chirps are transmitted in time division multiplexed (TDM) mode by transmitting sequentially through the two Tx antennas. This corresponds to an 8-element virtual array as shown in Figure 16. Therefore, for channel l, the channel information can be rewritten as: (Formula A7) TIFF2026009905000056.tif13122where, d l is the relative distance introduced by the virtual antenna l. θ is the azimuth angle of the target, as shown in Figure 16.

[0353] The phase of the channel information changes periodically in slow time due to the cyclical motion of breathing and heartbeat. Figure 18A shows a typical phase signal containing vital signs collected by the system.

[0354] In practice, target detection must precede vital sign detection, which is particularly difficult to achieve in indoor scenarios where there are various objects (e.g., walls, desks, metal objects, etc.) with strong EM wave reflections.

[0355] Range FFT and Digital Beamforming: In some embodiments, the channel information in the presence of static objects is: (Formula A8) TIFF2026009905000057.tif19148 where d0 is the distance between the object and the device, which remains constant in slow time.

[0356] The channel information corresponding to a reflecting object is a periodic signal in fast time, where the periodicity is related to distance as shown in equations A6 and A8. To determine the range information of a reflecting object, a Fast Fourier Transform (FFT) may be performed over fast time for each chirp, i.e., a range FFT may be performed, and the channel information is given by h r It may be written as (l,m), where r is the range tap index. The range tap corresponding to the reflective object will observe more energy compared to when there is no reflective object.

[0357] To further determine the azimuth angle of the reflecting object, digital beamforming is performed across all antenna elements for each range tap, and the channel information corresponding to range r and azimuth angle θ can be expressed as: (Formula A9) TIFF2026009905000058.tif1194where s H (θ) is the steering vector towards angle θ. In some embodiments of mmHRV, a Bartlett beamformer is applied, where the coefficients of the lth antenna are (Formula A10) TIFF2026009905000059.tif1977ε(m) is additive white Gaussian noise that is assumed to be independent and identically distributed (IID) for different range-azimuth bins. TIFF2026009905000060.tif11118 is the channel information vector at range tap r of all antenna elements. Therefore, for every slow time sample m, we have a channel information matrix h(r,θ) containing channel information at different location bins with range r and azimuth angle θ. Figure 17B shows the amplitude of the channel information in the range-azimuth plane.

[0358] Reflective Object Detector: In some embodiments, to locate a subject, it is first necessary to identify range-angle bins with reflective objects. The channel information of bins without reflective objects contains only noise. Therefore, the energy of the channel information of bins with reflective objects is greater than that of bins without reflective objects, as shown in Equations A6 and A8, respectively. However, it is difficult to find a universally predetermined threshold for target detection. According to the law of EM wave propagation, for the same reflective object, a shorter distance corresponds to greater reflected energy. Some mmHRV embodiments can utilize a constant false alarm rate (CFAR) detector, which can estimate the noise level by convolving the CFAR window (shown in FIG. 17A) with the channel information in the range-azimuth plane (shown in FIG. 17B). As shown in FIG. 17C, location bins with reflective objects are those whose energy exceeds the noise level. FIG. 17D shows an example of CFAR detection in the range domain, with the threshold indicated by the dashed line.

[0359] Subject Detector: In some embodiments, a reflective object detector can filter out empty taps, but cannot distinguish subjects from stationary reflective objects. Unlike static objects, the distance between the subject and the device changes over time due to motion (e.g., breathing and heartbeat), thus resulting in phase changes as shown in FIG. 18A. Therefore, the phase information of the candidate bins selected by the reflective object detector can be utilized to further filter out stationary reflective objects.

[0360] Figures 18A-18D show an example of a subject detector. The ground truth is that there are three subjects, one of which is sitting 1.5 m from the device at an azimuth angle of 0°, and the other two are sitting 1 m from the device at azimuth angles of 30° and -30°, respectively. Figure 18A shows phase information corresponding to the subject, Figure 18B shows phase information corresponding to static reflective objects, Figure 18C shows the results of the subject detector, with black dots corresponding to subjects, and Figure 18D shows the clustering results for each target.

[0361] When EM waves reflect off a subject, the phase changes over slow time due to modulation by human movement. Therefore, there is a large phase variance in the bins corresponding to the subject. However, for bins corresponding to static objects (e.g., desks, walls, etc.), the phase variance will be much smaller, as shown in Figures 18A and 18B. Therefore, in some embodiments of mmHRV, to filter out static objects, the variance of phase information over slow time can be examined, and the bins corresponding to the subject are those with phase variance above a certain threshold.

[0362] As shown in Figure 18C, considering the subject's volume, there will be more than one bin corresponding to the subject. To identify the target number, mmHRV utilizes a non-parametric clustering method, namely, the Density-Based Spatial Clustering for Applications with Noise (DBSCAN) algorithm, to cluster candidate bins without prior knowledge of the number of clusters in some embodiments. The clustering results are shown in Figure 18D. The representative of each cluster may be the bin with the best periodicity. For example, the bin with the highest peak relative to the first peak of the autocorrelation is selected, which corresponds to the bin with the highest SNR of the vital signs.

[0363] Heartbeat Extraction and HRV Estimation: In some embodiments, HRV estimation requires accurate estimation of the interbeat interval (IBI). Thus, mmHRV can extract the displacement changes caused by the heartbeat (also known as the heartbeat wave) from the composite displacement changes of the chest wall and detect the moment when the heartbeat occurs.

[0364] Heartbeat Extraction Algorithm: The phase information reflects the distance change caused by the vital signs. For simplicity, we can use the analog form of the signal directly, and the distance change of the human chest can be written as (Formula A11) TIFF2026009905000061.tif11108Here s m (t) denotes the distance change caused by the body movement. r (t) and s h (t) denotes the distance change caused by breathing and heartbeat, respectively. n(t) is the random phase offset introduced by noise, which is unrelated to the phase change caused by vital signs.

[0365] s r (t) and s h Both (t) are quasi-periodic signals whose period may vary slightly over time. Moreover, we can assume that body movements introduce a small number of oscillations, i.e., baseband signals. Therefore, the subject-related signals are sparse in the spectral domain, and we can reconstruct these signals with a small number of band-limited signals. For example, each component u k (t) is the central pulsation ω to be determined along with the decomposition k It is assumed to be compact around . Furthermore, the decomposition should simultaneously achieve spectral sparsity and data fidelity, which can be modeled as (Formula A12) TIFF2026009905000062.tif53158 where the first term evaluates the bandwidth of the analytical signal associated with each component, and the second term evaluates the data fidelity. K is the total number of resolved components, where TIFF2026009905000063.tif14127 is a set of all components and their center frequencies, respectively. α is a parameter used to balance bandwidth constraints and data fidelity.

[0366] Once the hyperparameters are known, the optimization problem in equation (A12) can be solved alternatively or iteratively until convergence occurs. k (t) and ω k To update uk, the subproblem can be written as follows: (Formula A13) TIFF2026009905000064.tif48159

[0367] Using Parseval's theorem, the problem can be rephrased as follows: (Formula A14) TIFF2026009905000065.tif43157 where, TIFF2026009905000066.tif1457 is u k is the Fourier transform of y(t) and y(t). After taking the integral over frequency and performing a change of variable, we can obtain the updated formula, where (Formula A15) TIFF2026009905000067.tif2293

[0368] center frequency ω k appears only in the bandwidth constraint, so the subproblem can be written as (Formula A16) TIFF2026009905000068.tif20149

[0369] As mentioned above, the optimum can be found in the Fourier domain, yielding: (Formula A17) TIFF2026009905000069.tif20124

[0370] The minimization of the quadratic problem above is (Formula A18) The file is TIFF2026009905000070.tif2475.

[0371] 19A-19B show the decomposition of a typical one-minute phase signal from an experiment, where the original phase information is decomposed into four components. FIG. 19A shows the decomposition result in the time domain, and FIG. 19B shows the corresponding spectrum of each component. In some embodiments, the first component 1901 reflects the subject's body movement, the second component 1902 is respiratory movement, and the third component 1903 is the heartbeat. Because noise has different vibration characteristics from vital signals, it falls into different modes, as does the signal decomposition residue 1904, as shown in FIGS. 19A-19B.

[0372] Once the hyperparameters are properly defined, the decomposition problem can be solved. However, it is difficult to predefine these hyperparameters in practical applications for heartbeat wave extraction. First, human motion is not always present, and human breathing may have a strong second harmonic component, making it even more difficult to determine the number of components. Furthermore, the hyperparameter α also affects the decomposition performance. Before discussing how to select the hyperparameters, the influence of the hyperparameters on the decomposition results is disclosed below.

[0373] If α is too small, i.e., the bandwidth constraint is too loose, and K is too small, mixing problems may occur, resulting in the two signals being merged into a single decomposed component. If K is too large, some of the decomposed components may contain noise. If α is too large, i.e., the bandwidth constraint is too tight, and K is too small, some target signals may be discarded by noise. If K is too large, some significant parts of the signal may be separated into two or more decomposed components.

[0374] In some embodiments of mmHRV, the number of components K and α can be adaptively changed for different data sets to accurately decompose the signal and obtain the component of interest, i.e., the heartbeat wave. A heuristic method for changing K and α as an iterative process to obtain an appropriate decomposition result is disclosed herein. Because the distance change caused by heartbeat is much smaller than the distance change caused by breathing and human movement, once the component corresponding to heartbeat is decomposed, the components corresponding to breathing and movement should also be decomposed in the same way, taking into account the data fidelity constraints in the objective function. Therefore, the algorithm terminates once the component corresponding to heartbeat is obtained.

[0375] HRV Estimation: In some embodiments, once the heart rate wave is extracted, the peaks in the heart rate wave can identify the exact time corresponding to each heart beat. For further accuracy, normalization may be performed before extracting the peaks.

[0376] In some embodiments, the envelope of the heartbeat wave is estimated by taking a moving average of the absolute value of the heartbeat component, as shown by the dashed line in FIG. 20A. To reduce noise, a moving average filter on the original heartbeat wave may be further performed. The normalized wave is the ratio of the filtered heartbeat wave to the estimated envelope. Therefore, the IBI can be derived by calculating the period of time between two adjacent heartbeats. FIG. 20B shows segments of the heartbeat wave and its ECG ground truth, where the dashed lines indicate the exact time of each heartbeat from a commercially available ECG sensor. The peaks of the normalized heartbeat wave align with the ground truth, and FIG. 20C shows the estimated IBI and ECG ground truth.

[0377] Further characteristics of HRV can be derived from the sequence of IBIs. In some embodiments of mmHRV, three metrics can be used to assess HRV: Root Mean Square of Successive Differences (RMSSD), which measures the change in successive IBIs, and can be calculated as follows: (Formula A19) TIFF2026009905000071.tif33140 where N IBI is the total number of IBIs measured. The standard deviation of all IBIs (SDRR) measures the variability of the IBIs and can be calculated as follows: (A20) TIFF2026009905000072.tif33113 where, TIFF2026009905000073.tif1216 is the empirical mean of the IBIs for each measurement. The metric pNN50 measures the proportion of consecutive IBIs that differ by more than 50 milliseconds (ms), which can be calculated by (Formula A21) TIFF2026009905000074.tif21141Here, 1{·} is the indicator function.

[0378] Experimental Evaluation: In some embodiments, a mmHRV system can be prototyped using a general-purpose millimeter-wave FMCW radar in a typical office space measuring 3.5 m x 3.2 m. By configuring two Tx and four Rx antennas in TDM-MIMO mode, the system can achieve a theoretical azimuth resolution of 15°. The field of view (FoV) is 100° in the horizontal plane with a radius of approximately 4 m, which is sufficient to cover a typical room. To obtain true heart rate signals, an ECG sensor is used to collect ground truth data simultaneously with mmHRV during the experiment. In total, 11 participants (6 males and 5 females) aged 20 to 60 years old are invited to conduct the experiment in both line-of-sight and non-line-of-sight scenarios. The experiments are conducted in a variety of settings, including different distances, angles of incidence, orientations, and obstructions between the subjects and the radar.

[0379] To further evaluate the performance of the proposed system, the mmHRV system can be compared with an HRV estimation technique using a bandpass filter bank (BPFB). Figure 21 shows the overall IBI estimation accuracy of the mmHRV and BPFB methods. The experiment included 11 participants, and 15 different experimental settings (e.g., different distances, incidence angles, orientations, and obstructions) were performed for each participant. As shown in Figure 21, BPFB yielded a median error of approximately 44 ms, with a 90th percentile error of approximately 200 ms. mmHRV achieved a median error of approximately 28 ms, with a 90th percentile error of 80 ms, outperforming BPFB by approximately 60%. To thoroughly evaluate the HRV estimation accuracy, Table I below shows the estimated HRV characteristics in terms of the average IBI, RMSSD, SDRR, and pNN50 for 11 participants with a user-device distance of approximately 1 m. We showed that mHRV can achieve a mean error of 3.89 ms for mean IBI, 6.43 ms for RMSSD, 6.44 ms for SDRR, and 2.52% for pNN50. Correspondingly, the mean estimation errors of BPFB are 15.33 ms for mean IBI, 41.94 ms for RMSSD, 32.59 ms for SDRR, and 12.17% for pNN50 estimation. Table 1 JPEG2026009905000075.jpg114169

[0380] 22 shows a flowchart of an example method 2200 for wireless vital signs monitoring according to some embodiments of the present disclosure. At operation 2202, a first wireless signal is transmitted over a wireless channel at a location. At operation 2204, a second wireless signal is received over the wireless channel, where the second wireless signal includes a reflection of the first wireless signal by at least one living thing having at least one repetitive movement within the location. At operation 2206, a time series of channel information (CI) for the wireless channel is obtained based on the second wireless signal, where each CI includes at least one of channel state information (CSI), channel impulse response (CIR), channel frequency response (CFR), or received signal strength index (RSSI). At operation 2208, for each of the at least one living thing, a vital signal representing all repetitive movements of the living thing is generated based on the plurality of TSCIs. At operation 2210, a heartbeat signal is extracted from the vital signal for each living thing. Heart rate variability is monitored for each living thing in the location based on the heart rate signal in act 2212. The order of the acts in Figure 22 may be varied according to various embodiments of the present disclosure.

[0381] In some embodiments, the wireless vital signs monitoring method includes steps s1 to s8 described below.

[0382] Step s1: Capture CSI using multiple transmit (Tx) antennas and multiple receive (Rx) antennas. Step s2: Apply beamforming to obtain directional CSI (e.g., CIR). Direction and distance may be associated with CSI. Step s3: Determine direction-of-interest (DoI) by detecting the presence of objects in each direction, including steps s3a and s3b performed for each direction.

[0383] Step s3a: Calculate the magnitude of the CSI (e.g., |h(theta, distance)| of the CIR) for each time instance. Then, time-average it over the time window. Step s3b: If the time-averaged magnitude response is greater than the threshold T1, it is determined that an object exists in the direction (i.e., the direction is the DoI), where the threshold T1 may be a two-dimensional CFAR filtering of |h| in the theta and distance directions.

[0384] Step s4: For each DoI (i.e., the direction in which the presence of an object is detected), perform motion detection by classifying the object into (a) a static object (e.g., furniture), (b) a stationary human (with breathing and heartbeat), and (c) random body movements including steps s4a to s4e.

[0385] Step s4a: Calculate the variance value (V) of the change over time of the phase of the CSI (e.g., the phase of h(theta, distance)) in a time window. In some embodiments, a larger phase variance means that the target is a living being with a heartbeat. Step s4b: If V is less than the threshold T2, classify the motion as a "static object". Step s4c: If V > T2, calculate the autocorrelation function (ACF) and find a significant feature point (e.g., the first peak) P1. Step s4d: If V > T2 and P1 > T3, classify the motion as a "stationary human". Step s4e: If V > T2 and P1 < T3, classify the motion as "random body movements". In some embodiments, a larger P1 means a more periodic feature.

[0386] Step s5: Determine the number of stationary subjects and their corresponding vital movements. This includes steps s5a to s5b.

[0387] Step s5a: Cluster the set of points of interest (PoIs) (i.e., (θ, distance) corresponding to the stationary humans in step s4), where the PoIs are clustered without prior knowledge of the number of clusters, i.e., non-parametric clustering. PoIs can be classified based on density-based techniques (e.g., DBSCAN) or based on distance (e.g., if the distance between two PoIs > the typical size of a human body, they belong to different clusters).

[0388] Step s5b: Generate vital movements corresponding to each subject. If two or more PoIs correspond to a subject, the corresponding movements can be combined, for example, by weighted averaging the phase measurements of the PoIs, or a dominant PoI can be identified and the vital movements associated with the dominant tap.

[0389] Step s6: For each subject, extract the heart rate signal by decomposing the vital signals using several band-limited signals, either by jointly optimizing the decomposition as in step s6a, or by successive decomposition as in step s6b.

[0390] Step s6a: Jointly optimize the decomposition of the modeled raw signals using steps s6a1, s6a2 and s6a3.

[0391] Step s6a1: Given an initial setting for the number of components K and parameter α to balance bandwidth constraints and data fidelity, alternately optimize the components and their center frequencies.

[0392] Step s6a2: Check whether there is a component corresponding to a heartbeat according to some features, and if the amplitude of the signal is in the range [T4, T5] and its center frequency is in the range [T6, T7], the component corresponds to a heartbeat wave. In some embodiments, the respiration / breathing signal can also be extracted in step s6a2.

[0393] If there is a decomposed component corresponding to the heartbeat in step s6a3, the heartbeat signal is normalized in step s7; if not, the number of components K and the value of the trade-off coefficient α are updated, and steps s6a1 to s6a3 are repeated.

[0394] Step s6b: Successively decompose the raw signal to obtain the heart rate wave using steps s6b1, s6b2, and s6b3.

[0395] Step s6b1: Process the raw signal by removing / suppressing the influence of the dominant (large magnitude) periodic signal (e.g., filtering the raw signal, or estimating the dominant periodic signal and subtracting it from the raw signal), where the dominant periodic signal can be estimated by operations on the raw signal (e.g., smoothing, low-pass filtering, spline interpolation, B-spline, cubic spline interpolation, polynomial fitting, polynomial fitting with an order adaptively selected based on distance / tap, etc.).

[0396] Step s6b2: Based on the processed raw signal, calculate the characteristics of the dominant periodic signal. The characteristics may be calculated based on a frequency transform, a trigonometric transform, a fast Fourier transform (FFT), a wavelet transform, an ACF, etc. Alternatively, the characteristics may be calculated by constrained optimization (e.g., minimization of an energy function subject to a smoothness constraint). The energy function may be the energy of the frequency (e.g., the energy of the FFT of the signal with the dominant component removed, where the signal may be a fused / clustered signal).

[0397] Step s6b3: Check whether the component corresponds to a heartbeat by some features. Here, if the amplitude of the signal is in the range [T4, T5] and its center frequency is in the range [T6, T7], the component corresponds to a heartbeat wave. If it corresponds to a heartbeat, normalize the heartbeat signal in step s7; if not, remove the component, and then repeat steps s6b2 and s6b3.

[0398] In some embodiments, other modal decomposition methods can be applied for step s6, where the modes can be viewed as frequency components, signals, etc., for example by ensemble empirical mode decomposition (EEMD). In some embodiments, instead of using phase information as input to extract the heartbeats, one can also rely on the CIR amplitude to extract the heartbeat signals / waves.

[0399] Step s7: Normalize the estimated heart rate wave by dividing it by the signal envelope, where the envelope can be estimated by operations on the raw signal (e.g., smoothing, low-pass filtering, spline interpolation, B-spline, cubic spline interpolation, polynomial fitting, and moving average).

[0400] Step s8: Identify the exact time of each heartbeat, and then calculate the interbeat intervals for estimating heart rate variability (HRV) and / or other statistics of the interbeat intervals, where the exact time of each heartbeat can be identified in several ways, for example, by identifying the peaks of the heartbeat wave, by identifying the zero-crossing points, or by finding some feature points after performing a continuous wavelet transform.

[0401] The following numbered sections provide examples of wireless vitals monitoring.

[0402] Clause B1. A system for wireless monitoring comprising: a transmitter configured to transmit a first wireless signal through a wireless channel at a location using N1 transmitting antennas; a receiver configured to receive a second wireless signal through the wireless channel using N2 receiving antennas, where N1 and N2 are positive integers, and the second wireless signal includes a reflection of the first wireless signal by at least one living thing having at least one repetitive movement at the location; and a processor configured to obtain a plurality of time series of channel information (TSCIs) for the wireless channel based on the second wireless signal, each of the plurality of TSCIs being associated with a respective transmitting antenna of the transmitter and a respective receiving antenna of the receiver; for each living thing of at least one living thing, generate a vital signal representing all repetitive movements of the living thing based on the plurality of TSCIs; extract a heart rate signal from the vital signals of each living thing; and monitor heart rate variability for each living thing at the location based on the heart rate signal.

[0403] Section B2. The system of section B1, wherein the at least one living thing includes a human or an animal, the first radio signal is carried by millimeter waves, and each object at the location has a position determined based on a plurality of spatial bins at the location, each of the plurality of spatial bins being determined by a direction and a distance range from the receiver, and each direction being associated with at least one of an angle, an azimuth angle, or an elevation angle.

[0404] Section B3. The system described in Section B2, wherein generating the vital signal for each living organism includes calculating beamforming based on the plurality of TSCIs and calculating a set of time series of directional channel information (CIs) each associated with a direction based on the beamforming.

[0405] Section B4. The system described in Section B3, wherein generating the vital signal for each living organism further includes: for each directional CI associated with each direction, calculating a CI amplitude for each time instance based on the directional CI for the respective direction and a distance range for obtaining CI amplitude over time; calculating a time average of the CI amplitude based on a time window; detecting the presence of an object in the distance range of the respective direction when the time average is greater than a first threshold; and determining a set of directions of interest (DoI), each including a direction in which the presence of an object is detected.

[0406] Section B5. The system of section B4, wherein the first threshold is adaptively determined to filter the CI amplitude at the distance range in the respective direction based on a two-dimensional constant false alarm rate (CFAR).

[0407] Section B6. The system described in Section B4, wherein generating the vital signal for each living organism further includes: for each DoI and each distance range of the set of DoIs, calculating a phase variance of a directional CI associated with the DoI over time within a time window; if the phase variance is less than a second threshold, classifying the object detected at the distance range and DoI as a static object without repetitive motion; and if the phase variance is greater than or equal to the second threshold, classifying the object detected at the distance range and DoI as a living organism with repetitive motion.

[0408] Section B7. The system described in Section B6, wherein generating the vital signal for each living thing further includes: determining a plur...

Claims

1. a transmitter configured to transmit a first wireless signal over a wireless channel at the location using N1 transmit antennas; a receiver configured to receive a second wireless signal over the wireless channel using N2 receive antennas; 1. A processor, comprising: obtaining time series of channel information (TSCI) for a plurality of the wireless channels based on the second wireless signal, each of the plurality of TSCIs being associated with a respective transmit antenna of the transmitter and a respective receive antenna of the receiver; generating, for each organism of at least one organism, a vital signal representative of all repetitive movements of the organism based on the plurality of TSCIs; extracting a heart rate signal from the vital signals of each living being; and a processor configured to: monitor, for each organism at the location, heart rate variability based on the heart rate signal; wherein N1 and N2 are positive integers, and the second wireless signal comprises a reflection of the first wireless signal by the at least one living thing performing at least one repetitive movement at the location.

2. 10. The system of claim 1, the transmitter and the receiver are physically coupled to each other; the at least one living organism comprises a human or an animal; the first radio signal is carried by millimeter waves; each object at the location has a location determined based on a plurality of spatial bins at the location; each of the plurality of spatial bins is determined by a direction and a range originating from the receiver; A system in which each direction is associated with at least one of an angle, an azimuth angle, or an elevation angle.

3. 3. The system of claim 2, wherein generating the vital signal for each living being comprises: calculating beamforming based on the plurality of TSCIs; and calculating a set of time series of directional channel information (CI) each associated with a direction based on the beamforming.

4. 4. The system of claim 3, wherein generating the vital signal for each living being comprises: For each directional CI associated with each direction, Calculating, for each time instance, a distance range for obtaining a CI amplitude based on the directional CI for each direction and a CI amplitude over time; calculating a time average of the CI amplitude based on a time window; detecting the presence of an object in the distance range in each direction when the time average is greater than a first threshold; determining a set of directions of interest (DoI), each of which includes a direction in which the presence of the object was detected.

5. 5. The system of claim 4, wherein the first threshold is adaptively determined to filter the CI amplitude at the distance range in the respective direction based on a two-dimensional constant false alarm rate (CFAR).

6. 5. The system of claim 4, wherein generating the vital signal for each living organism comprises: For each DoI and each distance range in the set of DoIs, Calculating a phase variance of a directional CI associated with said DoI over time in a time window; classifying the object detected in the distance range and DoI as a static object without repetitive motion when the phase variance is less than a second threshold; and classifying the object detected in the distance range and DoI as a living thing with repetitive motion when the phase variance is greater than or equal to the second threshold.

7. 7. The system of claim 6, wherein generating the vital signal for each living organism comprises: determining a plurality of target spatial bins for each detected organism, each of the plurality of target spatial bins being determined by a target DoI and a target distance range; For each target spatial bin, calculating an autocorrelation function based on the directional CI associated with the spatial bin of the target; determining a first peak of the autocorrelation function; classifying the detected creature movement in the target spatial bin as repetitive movement when the first peak is greater than a third threshold; classifying the detected creature movement in the target spatial bin as random body movement when the first peak is less than or equal to a third threshold.

8. 8. The system of claim 7, wherein generating the vital signal for each living organism comprises: Calculating a set of points of interest (PoIs), each PoI in the set of PoIs being associated with a living thing detected at the PoI and a target spatial bin at the location associated with a repetitive movement of the living thing detected at the PoI; clustering the set of PoIs to generate at least one PoI cluster out of a total number of clusters; the set of PoIs is clustered without prior knowledge of the total number of clusters after clustering; clustering, wherein the set of PoIs is clustered based on at least one of a density associated with the set of PoIs, a distance between any two PoIs in the set of PoIs, or a threshold associated with the size of an organism; determining an abundance of the target organism at the location based on the total number of clusters.

9. 9. The system of claim 8, wherein generating the vital signal for each living organism comprises: For each of the at least one PoI cluster, combining PoIs in the PoI cluster based on at least one of a weighted average of CI phases measured at PoIs or a dominant PoI having the highest peak among the first peaks of the autocorrelation function associated with the PoIs to generate a combined PoI; The system further includes generating a vital signal for the target organism corresponding to the PoI cluster based on the CI phase signal corresponding to the combined PoI, wherein the CI phase signal is associated with all repetitive movements of the target organism.

10. 10. The system of claim 9, wherein extracting a heart rate signal from the vital signals comprises: The system includes, for each organism, decomposing the CI phase signal associated with the organism to generate a cardiac signal based on at least one of a joint optimization of the decomposition of the CI phase signal or a continuous decomposition of the CI phase signal.

11. 11. The system of claim 10, wherein the joint optimization comprises: determining a number K representing an amount of possible signal components of the CI phase signal, K being equal to or greater than the amount of organisms at the location; Determining a trade-off factor to balance bandwidth constraints and data fidelity; iteratively optimizing an objective function that simultaneously maximizes spectral sparsity and data fidelity of the CI phase signal based on the trade-off coefficient, the K signal components of the CI phase signal, and the center frequencies of the K signal components, until the objective function converges; and simultaneously generating K decomposed components of the CI phase signal based on the iterative optimization.

12. 12. The system of claim 11, wherein the joint optimization comprises: The system further includes determining whether the K decomposed components include a heartbeat component having an amplitude within a first value range and a center frequency within a second value range, each of the first value range and the second value range being predetermined based on heartbeat statistics.

13. 13. The system of claim 12, wherein the joint optimization comprises: When the K decomposed components include a heartbeat component, estimating an envelope of the cardiac component based on at least one of smoothing, low-pass filtering, spline interpolation, B-spline, cubic spline interpolation, polynomial fitting, or moving average; normalizing the heart rate component by dividing the heart rate component by the envelope of the heart rate component to generate a normalized heart rate signal for the organism.

14. 13. The system of claim 12, wherein the joint optimization comprises: If there is no heartbeat component in the K decomposed components, updating the number K to generate an updated number K' representing an updated amount of possible signal components of the CI phase signal; updating the trade-off factor to generate an updated trade-off factor for balancing bandwidth constraints and data fidelity; The system further includes iteratively optimizing an objective function that simultaneously maximizes spectral sparsity and data fidelity of the CI phase signal based on the updated trade-off coefficient, the K' signal components of the CI phase signal, and the center frequencies of the K' signal components, until the objective function converges, thereby generating K' decomposed components of the CI phase signal.

15. 11. The system of claim 10, wherein the continuous decomposition comprises: estimating a dominant component of the CI phase signal based on at least one of smoothing, low-pass filtering, spline interpolation, or polynomial fitting; removing the dominant component from the CI phase signal to generate a processed CI phase signal; calculating a characteristic of a second dominant component of the CI phase signal based on the processed CI phase signal using at least one of a frequency transform, a trigonometric transform, a fast Fourier transform (FFT), or a wavelet transform; and determining, based on the characteristic, whether the second dominant component is a heartbeat component having an amplitude within a first range of values ​​and a center frequency within a second range of values, each of the first range of values ​​and the second range of values ​​associated with a heartbeat.

16. 16. The system of claim 15, wherein the successive decomposition comprises: When the second dominant component is a heartbeat component, estimating an envelope of the cardiac component based on at least one of smoothing, low-pass filtering, spline interpolation, B-spline, cubic spline interpolation, polynomial fitting, or moving average; normalizing the heart rate component by dividing the heart rate component by the envelope of the heart rate component to generate a normalized heart rate signal for the organism.

17. 16. The system of claim 15, wherein the successive decomposition comprises: When the second dominant component is not a heartbeat component, removing the second dominant component from the CI phase signal to generate an additional processed CI phase signal; calculating additional characteristics of a next dominant component of the CI phase signal based on the additional processed CI phase signal; determining, based on the additional characteristic, whether the next dominant component is a cardiac component having an amplitude within the first range of values ​​and a center frequency within the second range of values.

18. 11. The system of claim 10, wherein monitoring the heart rate variability comprises: For each organism, a cardiac cycle time for the organism at each time instance; calculating a cardiac time based on at least one of identifying peaks in the cardiac signal, identifying zero crossings in the cardiac signal, and performing a continuous wavelet transform on the cardiac signal; calculating a plurality of interbeat intervals based on the heartbeat times; estimating the heart rate variability for the organism based on statistics of the interbeat intervals.

19. 3. The system according to claim 2, wherein extracting a heart rate signal from the vital signals comprises, for each living being: Decomposition of the vital signal based on frequency components of the vital signal; or generating the cardiac signal based on at least one of decompositions of a CI amplitude signal associated with the organism.

20. 1. A method for a wireless monitoring system, comprising: transmitting a first wireless signal over a wireless channel of the location; receiving a second wireless signal over the wireless channel, the second wireless signal comprising reflections of the first wireless signal by a plurality of persons at the location; Obtaining a time series of channel information (TSCI) of the wireless channel based on the second wireless signal, where each CI includes at least one of a channel state information (CSI), a channel impulse response (CIR), a channel frequency response (CFR), or a received signal strength index (RSSI); generating, for each of the plurality of humans, a vital signal representative of all repetitive movements of the human based on the TSCI; extracting a heart rate signal from the vital signals of each person; and simultaneously monitoring heart rate variability for each of the plurality of humans based on the heart rate signals.

Images