SYSTEMS AND METHODS FOR PERSONAL STATUS DIFFERENTIATION

Abstract

A method for determining a person's status using wireless signals involves annotating selected channel state information segments with a class indicating the person's status based on their gait and at least some of a variety of biometric characteristics. The method also involves identifying the person's status based on their gait and at least some of their biometric characteristics, determining one or more settings of a machine corresponding to the person's status, and controlling, by a computer system and using the one or more settings, the operation of the machine in response to the person's status being detected.

Description

TECHNICAL AREA

[0001] The present disclosure relates to person status identification and, in particular, person status differentiation using wireless signals. BACKGROUND OF THE INVENTION

[0002] Person status identification is the process of recognizing and verifying a person's status, for example, whether they are a child or an adult. Applications in this area range from security and access control to authentication in various domains. Methods such as facial recognition, fingerprint reading, voice authentication, and retinal verification offer a biometric approach to using biometric data to verify a person's status. Furthermore, these methods utilize physical interactions between the system and the individual. SUMMARY

[0003] One aspect of the disclosed embodiments includes a method for determining a person's status using wireless signals. The method involves collecting, at a wireless receiver, channel state information from received packets transmitted by a wireless transmitter; annotating, using the computer system, selected channel state information segments with a class indicating the person's status based on a gait and at least some of a variety of biometric characteristics of the person extracted from the channel state information;Identifying, using a machine learning model, the status of the person based on their gait and at least some of the person's multiple biometric characteristics, wherein the machine learning model is trained using classifier training and training data comprising information from the selected channel state information segments and gait and biometric information from persons with different statuses; determining, using the computer system and the machine learning model and based on the person's status, one or more settings of a machine located at a site comprising the wireless receiver and the wireless transmitter, wherein the one or more settings correspond to the person's status;and control, by the computer system and using one or more settings, of the operation of the machine in response to the detection of the person's status;

[0004] Another aspect of the disclosed embodiments includes a system for determining a person's status using wireless signals. The system includes: a wireless receiver configured to receive packets transmitted by a wireless transmitter and further configured to collect channel state information from the packets; and a computer system associated with the wireless receiver, the computer system being configured to: annotate the selected channel state information segments with a class indicating a person's status, based on a gait and at least some of a variety of the person's biometric characteristics extracted from the channel state information;Identifying, using a machine learning model, the status of the person based on their gait and at least some of the person's multiple biometric characteristics, wherein the machine learning model is trained using classifier training and training data comprising information from the selected channel state information segments and gait and biometric information from persons with different statuses; determining, using the machine learning model and based on the person's status, one or more settings of a machine located at a site comprising the wireless receiver and wireless transmitter, wherein the one or more settings correspond to the person's status; and controlling, using the one or more settings, the operation of the machine in response to the detection of the person's status.

[0005] Another aspect of the disclosed embodiments involves a non-volatile, computer-readable medium that stores instructions which, when executed by a computer system, cause the computer system to perform operations that include: annotating the selected channel state information segments with a class indicating a person's status based on their gait and at least some of a variety of biometric features of the person extracted from the channel state information; and identifying, using a machine learning model, the person's status based on their gait and at least some of the variety of biometric features of the person, wherein the status indicates whether the person is a child or an adult.Determine, using the machine learning model and based on the person's status, one or more settings of a machine located at a site containing the wireless receiver, wherein the one or more settings correspond to the person's status; and control, using the one or more settings, the operation of the machine in response to the detection of the person's status. BRIEF DESCRIPTION OF THE DRAWINGS Fig. Figure 1 shows a System 100 for training a neural network. Fig. Figure 2 shows a system for implementing the machine learning models. Fig. Figure 3A shows an example of an embodiment of a system that uses information from wireless signals for person status identification. Fig. Figure 3B shows an example of information that can be extracted from wireless signals for use in an embodiment of a method for determining a person status identification. Fig. Figure 3C shows a workflow diagram for an embodiment of a method according to the disclosure. Fig. Figure 4A shows a workflow diagram and signal characteristics for amplitude preprocessing of channel state information (CSI) for an embodiment of the disclosed method. Fig. Figure 4B shows a workflow diagram and signal characteristics for a phase preprocessing of channel state information (CSI) for an embodiment of the disclosed method. Fig. Figure 4C shows a neural network used in one embodiment of the disclosure. Fig. 4D is a workflow diagram that represents an embodiment of a method for performing an inference according to the disclosure. Fig. 4E is a workflow diagram that represents an embodiment of a method for determining a person's status according to the disclosure. Fig. Figure 5 shows a schematic diagram of an interaction between a computer-controlled machine 510 and a control system 512. Fig. Figure 6 shows a schematic diagram of the tax system of Fig. 1, which is configured to control a vehicle, which may be a semi-autonomous vehicle or a semi-autonomous robot. Fig. Figure 7 shows a schematic diagram of the tax system of Fig. 1, which is configured to control a manufacturing machine, such as a punch cutter, cutter or pistol drill, or a manufacturing system, such as part of a production line. Fig. Figure 8 shows a schematic diagram of the tax system of Fig. 1, which is configured to control a power tool, such as an electric drill or screwdriver, which has at least a semi-autonomous mode. Fig. Figure 9 shows a schematic diagram of the tax system of Fig. 1, which is configured to control an automated personal assistant. Fig. Figure 10 shows a schematic diagram of the tax system of Fig. 1, which is configured to control a monitoring system, such as a control access system or a surveillance system. Fig. Figure 11 shows a schematic diagram of the tax system of Fig. 1, which is configured to control an imaging system, for example, an MM device, an X-ray imaging device, or an ultrasound device. DETAILED DESCRIPTION

[0006] Embodiments of the present disclosure are described herein. It is understood, however, that the disclosed embodiments are merely examples and that other embodiments may take different and alternative forms. The figures are not necessarily to scale; some features may be enlarged or reduced to show details of certain components. Therefore, the specific structural and functional details disclosed herein are not to be interpreted as limiting, but merely as a representative basis for teaching a person skilled in the art how to use the embodiments in various ways.As the person skilled in the art will understand, various features illustrated and described with reference to any of the figures can be combined with features illustrated in one or more other figures to create embodiments not explicitly illustrated or described. The combinations of illustrated features provide representative embodiments for a typical application. However, various combinations and modifications of the features, consistent with the teachings of this disclosure, may be desirable for certain applications or implementations.

[0007] “A”, “an”, and “the”, as used here, refer to both singular and plural referents unless the context clearly indicates otherwise. For example, “a processor” programmed to perform various functions refers to a processor programmed to perform each individual function, or to more than one processor programmed together to perform each of the different functions.

[0008] Personal identification, or PID, is the process of uniquely identifying and verifying an individual's identity. It encompasses applications ranging from security and access control to authentication across various domains. Typically, PID utilizes diverse data sources, such as biometrics, facial recognition, audio data, fingerprinting, voice authentication, and ophthalmo-retinal verification technologies, to positively identify an individual. While these technologies offer a robust approach to verifying human identity, they can be highly invasive to individual privacy. The risk of data theft, misuse, and spoofing remains a concern. Furthermore, users may experience some inconvenience while interacting with these authentication systems, often requiring them to be in close proximity to the system.

[0009] Compared to these methods, wireless signal-based personal identification (PID) methods offer seamless convenience while maintaining security. These methods operate over a wide area by utilizing multi-path radio signals that propagate in all directions. This allows for broad range and signal coverage for human identification, even when the user is not within a small detection zone. Additionally, wireless signal-based PID systems can preserve privacy because the signals, which contain identifiable signatures, are affected by the environment and may be limited to the detection zone.While traditional identifiable information such as fingerprint data remains static once captured (which can increase the risk of theft), wireless signals vary across space and time, meaning that wireless biometric data may not be useful for identifying the same person in a different environment. Therefore, wireless signal-based identification systems can be very convenient and secure.

[0010] Wireless PID systems can also be useful for personalized content delivery. Previous systems may not be able to automatically detect human presence and identify a nearby person, instead delivering content based on either proactive human requests or events at the time of day. While such a system can be useful for certain daily routines, a change in a routine could render the system unhelpful for that user. However, the use of wireless PID according to the present disclosure for intelligent assistance systems can enable the detection and identification of human presence in order to deliver personalized content once a person has been identified.For example, a smart home appliance, such as a smart coffee maker, could automatically detect the presence of a specific person and dispense coffee brewed according to their personal preferences. In another example, a smart thermostat system could set the desired temperature based on the presence of a particular person. Accordingly, this PID mechanism can enable learning-based household appliances. Beyond household appliances, wireless PID can enable the use of smart machines in industrial settings, workplaces, and a variety of other applications where a machine can be controlled based on personal usage patterns and / or preferences.

[0011] In addition to the above, the wireless PID systems of this disclosure can also provide a crude identification and perform functions based thereon. For example, a wireless PID system according to the disclosure can determine whether a detected person is an adult or a child and can restrict access to some household appliances (e.g., a stove, an oven, etc.) based on this.

[0012] The wireless PID systems and methods of this disclosure protect privacy, are non-invasive, and can be implemented in a variety of embodiments. In one embodiment, the method can utilize an existing Wi-Fi transmitter (e.g., a home Wi-Fi router) and a wireless receiver configured to receive wireless signals transmitted by the transmitter. The wireless receiver can be integrated, for example, into a household appliance or security system and can continuously monitor wireless signals transmitted by the transmitter. These wireless signals can be reflected and / or partially absorbed by a person near a path between the transmitter and the receiver. By receiving the wireless signals, the receiver can gather information regarding gait and other movement signatures / characteristics (e.g., walking patterns).Since human gait can be unique to each individual due to differences in body shape, muscle, and fat tissue, the systems and methods disclosed herein can learn to identify individuals, particularly through the use of a neural network. The neural network can learn other biometric features, including height, body shape, and so on. Furthermore, the systems and methods disclosed herein can detect activity and patterns of these features for identified individuals. Using artificial intelligence and neural networks, the systems and methods disclosed herein can utilize Wi-Fi channel state information (CSI) features for different individuals to solve a multi-class classification task. The methodology can also be extended to other wireless technologies, such as Bluetooth, ultra-wideband (UWB), LTE, 5G, 6G, and so forth, and is therefore not limited exclusively to Wi-Fi implementations.It can also work with multiple antennas.

[0013] In some embodiments, the systems and methods described herein may be configured to determine a person's status using wireless signals. For example, existing household appliances (e.g., kitchen stoves, conventional ovens, and / or the like) do not have the ability to detect whether a person interacting with such an appliance has a first status, such as a child status (e.g., younger than a threshold, such as 10 years old, 12 years old, 18 years old, or any suitable threshold), or a second status, such as an adult status (e.g., older than the threshold). As a result, a child could turn on a stove or open a hot oven, potentially causing injury and / or property damage.Accordingly, the systems and procedures described herein may be configured to use wireless signals to determine whether a person interacting with a machine (e.g., a home alarm, a vehicle, a power tool, and / or any machine described herein or any other suitable machine) is a child, and in response to determining that the person is a child, to activate a feature or function, such as an interlock, to prevent injury and / or property damage (e.g., when the interlock is activated, the device will not allow a child to turn on a stove or open a hot oven).Additionally or alternatively, the systems and procedures described herein may be configured to allow adults or machine users to configure a dishwasher, washing machine, dryer, or other appliance so that the systems and procedures can prevent access to, use of, and / or the like from the machine based on the person's status (e.g., if the person's status indicates that the person is a child). The systems and procedures described herein may also be configured to allow such machines to be configurable to restrict access to and / or use based on the person's status.

[0014] Additionally or alternatively, the systems and procedures described here can be configured to provide restricted access to personal assistance devices (e.g., smart speakers, smart monitors, and / or the like) based on the person's status, so that a person with a status indicating that the person is a child can be prevented from accessing content (e.g., specific content) on such personal assistance devices.

[0015] In some embodiments, the systems and methods described herein can be configured to interact with a machine incorporating a wireless chipset, as described herein, which receives wireless packets and extracts channel state information for each packet. The systems and methods described herein can be configured to use the wireless signals to determine a person's body shape, gait, muscle and fat tissue reflections, and / or the like, within a space assigned to the machine. The systems and methods described herein can be configured to differentiate between body shape, gait, muscle tissue, fat tissue, and the like for a child and an adult in order to determine the person's status (e.g., child status or adult status).

[0016] In some embodiments, the systems and methods described herein may be configured to use Wi-Fi or other suitable wireless technology, including but not limited to Bluetooth, Ultra Wideband (UWB), 5G, 6G and / or other suitable wireless technologies, to determine the status of the person.

[0017] In addition to or as an alternative to the characteristics described here for determining a person's identity, the systems and procedures described here can be configured to annotate each time segment (e.g., 2 seconds of data) with a class, such as child or adult, after data collection and cleaning phases (e.g., as described here). The systems and procedures described here can be configured to set the total number of classes to two. The classifier can be configured to learn how to classify these classes based on the classifiers described here.

[0018] In some embodiments, the systems and methods described here can be configured to set three classes (e.g., child, adult, or empty space). The classifier can classify these classes based on the classifiers described here.

[0019] The systems and procedures described here can be configured to collect annotations from the user as an initial setup. They can also be configured to use the annotated data to train the machine learning model. Alternatively, or in addition, the user can choose not to provide annotations, and the systems and procedures described here can be configured to provide annotations during product development when a large dataset with such annotations is collected from various environments and the classifiers are trained using the annotated data. The user then uses the configured classifiers for inference.

[0020] The systems and methods described herein can be configured to provide an action upon successful detection of a child interacting with the machine (e.g., preventing a child from turning on the stove, preventing a child from opening an oven door based on the oven's heat, and / or the like). In some embodiments, the systems and methods described herein can be configured to provide the action based on one or more specific machine settings (e.g., where the one or more settings are configured by the user, who defines the action to be taken in response to determining that the person within the machine's space is a child).

[0021] Fig. Figure 1 shows a System 100 for training a neural network, such as a deep neural network. The neural network or deep neural networks shown and described are merely examples of the types of machine learning networks or neural networks that can be used. The System 100 may include an input interface for accessing training data 102 for the neural network. For example, as shown in Fig. As shown in Figure 1, the input interface is formed by a data storage interface 104, which can access the training data 102 from a data storage device 106. For example, the data storage interface 104 can be a storage interface or a persistent storage interface, such as a hard disk or SSD interface, but also a personal, local, or wide-area network interface, such as a Bluetooth, Zigbee, or Wi-Fi interface, or an Ethernet or fiber optic interface. The data storage device 106 can be internal data storage of the system 100, such as a hard disk or SSD, but also external data storage, such as network-accessible data storage.

[0022] In some embodiments, the data storage 106 may further comprise a data representation 108 of an untrained version of the neural network, which the system 100 can access from the data storage 106. It is understood, however, that the training data 102 and the data representation 108 of the untrained neural network can each also be accessed from another data storage, e.g., via another subsystem of the data storage interface 104. Each subsystem may be of a type such as described above for the data storage interface 104. In other embodiments, the data representation 108 of the untrained neural network may be generated internally by the system 100 based on design parameters for the neural network and therefore does not need to be explicitly stored on the data storage 106.

[0023] System 100 can further include a processor subsystem 110, which can be configured to provide an iterative function during the operation of System 100 as a replacement for a stack of layers of the neural network to be trained. Here, the respective layers of the stack being replaced can have shared weights and can receive as input an output from a previous layer, or, for the first layer of the stack, an initial activation and part of the stack's input. Processor subsystem 110 can also be configured to iteratively train the neural network using the training data 102. Here, an iteration of the training by processor subsystem 110 can include a forward propagation part and a backward propagation part.

[0024] The processor subsystem 110 can be configured to perform the forward propagation part by determining, among other operations that define the forward propagation part that can be performed, an equilibrium point of the iterative function at which the iterative function converges to a fixed point, wherein determining the equilibrium point involves using a numerical root-finding algorithm to find a root solution for the iterative function minus its input, and by providing the equilibrium point as a substitute for an output of the stack of layers in the neural network.

[0025] System 100 can further include an output interface for outputting a data representation 112 of the trained neural network, whereby this data can also be referred to as trained model data 112. For example, as also in Fig. As shown in Figure 1, the output interface is formed by the data storage interface 104, wherein the interface in these embodiments is an input / output (“I / O”) interface through which the trained model data 112 can be stored in the data storage 106. For example, the data representation 108, which defines the “untrained” neural network, can be at least partially replaced during or after training by the data representation 112 of the trained neural network by adjusting the parameters of the neural network, such as weights, hyperparameters, and other types of neural network parameters, to reflect the training on the training data 102.

[0026] This is also in Fig. 1 is represented by the reference symbols 108 and 112, which refer to the same data set on the data storage 106. In other embodiments, the data representation 112 can be stored separately from the data representation 108, which defines the “untrained” neural network. In some embodiments, the output interface can be separate from the data storage interface 104, but can generally be of a type such as described above for the data storage interface 104.

[0027] In various embodiments, the neural network training system can be implemented within a personal identification system using wireless signals (e.g., Wi-Fi) received by a wireless receiver. The data (e.g., CSI) obtained from the wireless signals can be used to determine gait and / or other movement characteristics to identify a specific individual. Embodiments in which the data can also be used for cruder identification (e.g., to distinguish between an adult and a child) using the neural network training system are also possible and are under consideration. The system can be implemented in a household appliance, industrial equipment, or any other suitable environment.Based on the training, the neural network can be used to determine the patterns of the specific person and can adjust its operation accordingly.

[0028] Fig. Figure 2 shows a System 200 for implementing the machine learning models described herein, for example, the deep neural networks used to perform personal identification using data obtained from received wireless signals, as described above and in more detail below. Other types of machine learning models may be used, and the DNNs described herein are not the only types of machine learning models that may be used in the system of this disclosure. For example, if the input image contains an ordered sequence of pixels (after converting CSI values ​​into pixels in an image), a CNN may be used. The System 200 may be implemented to perform one or more of the image recognition stages described herein. The System 200 may include at least one Computing System 202.

[0029] The computing system 202 can include at least one processor 204 operatively connected to a memory unit 208. The processor 204 can include one or more integrated circuits that implement the functionality of a central processing unit (CPU) 206. The CPU 206 can be a commercially available processing unit that implements an instruction set such as one of the x86, ARM, Power, or MIPS instruction set families. During operation, the CPU 206 can execute stored program instructions retrieved from the memory unit 208. The stored program instructions can include software that controls the operation of the CPU 206 to perform the operation described herein. In some examples, the processor 204 can be a system-on-a-chip (SoC) that integrates the functionality of the CPU 206, the memory unit 208, a network interface, and input / output interfaces into a single integrated device.The computer system 202 can implement an operating system to manage various aspects of the operation. While a processor 204, a CPU 206, and a memory 208 are included in... Fig. As shown in the 2 examples, of course more than one of each can be used in an overall system.

[0030] The memory unit 208 can include volatile and non-volatile memory for storing instructions and data. The non-volatile memory can include solid-state memory such as NAND flash memory, magnetic and optical storage media, or any other suitable data storage device that retains data when the computing system 202 is disabled or loses electrical power. The volatile memory can include static and dynamic random-access memory (RAM) that stores program instructions and data. For example, the memory unit 208 can store a machine learning model 210 or an algorithm, a training dataset 212 for the machine learning model 210, and a raw source dataset 216.

[0031] The computing system 202 may include a network interface device 222 configured to provide communication with external systems and devices. For example, the network interface device 222 may include a wired and / or wireless Ethernet interface, as defined by the 802.11 standard family of the Institute of Electrical and Electronics Engineers (IEEE). The network interface device 222 may include a cellular communication interface for communicating with a cellular network (e.g., 3G, 4G, 5G). The network interface device 222 may also be configured to provide a communication interface to an external network 224 or a cloud.

[0032] The external network 224 can be referred to as the World Wide Web or the Internet. The external network 224 can establish a standard communication protocol between computing devices. The external network 224 enables the easy exchange of information and data between computing devices and networks. One or more servers 230 can communicate with the external network 224.

[0033] The computing system 202 may include an input / output (I / O) interface 220, which can be configured to provide digital and / or analog inputs and outputs. The I / O interface 220 is used to transfer information between internal memory and external input and / or output devices (e.g., HMI devices). The I / O interface 220 may include associated circuitry or bus networks to transfer information to or between the processor(s) and memory. For example, the I / O interface 220 may include digital I / O logic lines that can be read or set by the processor(s), handshake lines to monitor data transmission over the I / O lines, timing and counter devices, and other structures known to provide such functions. Examples of input devices include a keyboard, mouse, sensors, etc.Examples of output devices include monitors, printers, speakers, etc. The I / O interface 220 can include additional serial interfaces for communicating with external devices (e.g., a Universal Serial Bus (USB) interface). The I / O interface 220 can be described as an input interface (in which it transmits data from an external input, such as a sensor) or an output interface (in which it transmits data to an external output, such as a display).

[0034] The computing system 202 may include a human-machine interface (HMI) device 218, which may include any device that enables the system 200 to receive control inputs. Examples of input devices may include human interface inputs such as keyboards, mice, touchscreens, speech input devices, and other similar devices. The computing system 202 may include a display device 232. The computing system 202 may include hardware and software for outputting graphic and text information to the display device 232. The display device 232 may include an electronic display screen, projector, printer, or other suitable device for displaying information to a user or operator. The computing system 202 may further be configured to allow interaction with remote HMIs and remote display devices via the network interface device 222.

[0035] System 200 can be implemented using one or more computing systems. While the example represents a single computing system 202 that implements all of the described features, it is intended that different features and functions can be implemented separately and by multiple computing units communicating with each other. The specific system architecture chosen may depend on a variety of factors.

[0036] System 200 can implement a machine learning algorithm 210 configured to analyze the raw source dataset 216. The raw source dataset 216 can contain raw or unprocessed sensor data that may be representative of an input dataset for a machine learning system. The raw source dataset 216 can include video, video segments, images, text-based information, audio or human speech, time-series data (e.g., a pressure sensor signal over time), raw or partially processed sensor data (e.g., a radar map of objects), and wireless signals in the form of CSI, RSSI, or CIR.Furthermore, the raw source dataset can be input data derived from an associated sensor, such as a camera, lidar, radar, ultrasonic sensor, motion sensor, thermal imaging camera, wireless receivers, or any other type of sensor that produces associated data with spatial dimensions, where there is some notion of a "foreground" and a "background" within those spatial dimensions.

[0037] References to an input or an input "image" do not necessarily originate from a camera, but can come from any of the sensors listed above. Several different examples of inputs are given with reference to Fig. 6- Fig. 12 shown and described. In some examples, the machine learning algorithm 210 can be a neural network algorithm (e.g., a deep neural network) designed to perform a predetermined function.

[0038] The computer system 200 can store a training dataset 212 for the machine learning algorithm 210. The training dataset 212 can represent a set of previously constructed data for training the machine learning algorithm 210. The training dataset 212 can be used by the machine learning algorithm 210 to learn weighting factors associated with a neural network algorithm. The training dataset 212 can include a set of source data exhibiting corresponding results or outcomes that the machine learning algorithm 210 attempts to duplicate through the learning process.

[0039] The machine learning algorithm 210 can be operated in a learning mode using the training dataset 212 as input. The machine learning algorithm 210 can be executed over a number of iterations using the data from the training dataset 212. With each iteration, the machine learning algorithm 210 can update internal weighting factors based on the results obtained. For example, the machine learning algorithm 210 can compare output results (e.g., a reconstructed or augmented image, in the case where image data is the input) with those included in the training dataset 212. Since the training dataset 212 contains the expected results, the machine learning algorithm 210 can determine when performance is acceptable. After the machine learning algorithm 210 reaches a predetermined performance level (e.g.,Once 100% agreement with the results associated with the training dataset 212 is achieved, or convergence is reached, the machine learning algorithm 210 can be run using data not contained in the training dataset 212. It is understood that in this disclosure, "convergence" can mean that a set (e.g., predetermined) number of iterations has occurred, or that the remainder is sufficiently small (e.g., the change in the approximate probability across iterations is less than a threshold), or other convergence conditions. The trained machine learning algorithm 210 can be applied to new datasets to generate annotated data.

[0040] The machine learning algorithm 210 can be configured to identify a specific feature in the raw source data 216. The raw source data 216 can include a variety of instances or an input dataset for which supplemental results are desired. The machine learning algorithm 210 can be programmed to process the raw source data 216 to identify the presence of the specific features. The machine learning algorithm 210 can be configured to identify a feature in the raw source data 216 as a predetermined feature. The raw source data 216 can be derived from a variety of sources. For example, the raw source data 216 can be actual input data collected by a machine learning system. The raw source data 216 can be machine-generated for testing the system.For example, the raw source data could include 216 raw video images from a camera, wireless signals and / or the like.

[0041] Fig. Figure 3A shows an example of an embodiment of a system that uses information from wireless signals to identify the status of a person. In the example shown, wireless signals 325A, 325B, and 325C are sent from the transmitter 302 to the receiver 304. The wireless transmitter 302 can be a Wi-Fi router in a home in various embodiments, although the scope of the disclosure is not limited to Wi-Fi implementations and thus considers others (e.g., Bluetooth). Fig. 3A represents a home environment, the disclosure is similarly not limited in this way and thus the various systems and procedures disclosed herein can be implemented in a variety of environments.

[0042] In the embodiment shown, the receiver 304 is embedded in a household appliance (e.g., a coffee machine) that can be operated as an Internet of Things (IoT) device. More generally, the receiver can be part of virtually any type of device or equipment capable of receiving and processing wireless signals transmitted by the transmitter 302. Both the transmitter 302 and the receiver 304 are located in a space (e.g., a room in a house) that may include other static surfaces 340, such as cabinets, walls, floors, ceilings, furniture, and so on. In this example, a person 310 is located along a path between the transmitter 302 and the receiver 304 and may attenuate or block wireless signals along that particular path.

[0043] The wireless signals 325A, 325B and 325C transmitted by the transmitter are subject to multipath propagation, as shown in Fig. Figure 3A shows that signals can be reflected by various static surfaces 340 before being detected by the receiver 304. Due to the different paths and thus the propagation length, parts of wireless signals transmitted at a given time may be received by the receiver 304 at different times. Accordingly, the receiver 304 can use beamforming techniques to combine received signals as they are received at different angles (via different paths) to improve the received signal strength.

[0044] In the illustrated embodiment, the receiver 304 can utilize multipath propagation and environmental knowledge to detect the presence and movement of person 310. Since person 310 can attenuate or block some wireless signals, their presence and movement can be detected by the receiver 304. Upon detecting the presence of person 310 (and more generally, any person), the receiver 304 can perform both coarse-grained and fine-grained identification using various artificial intelligence / machine learning (AI / ML) techniques. For example, coarse-grained identification can determine whether the detected person is an adult or a child and can also estimate their approximate height.For fine-grained identification, the receiver 304 can determine the presence of a specific person 310 based on factors such as gait, other movement signatures, and so on, and identify them.

[0045] Using the AI / ML techniques of this disclosure, a model can be subjected to multi-classifier training to identify different individuals who can perform different actions at the location containing both the sender 302 and the receiver 304. Using the AI / ML techniques, the receiver 304 can also detect usage patterns for the different individuals and perform additional actions based on these patterns. For example, if the receiver 304 is embedded in a coffee machine, it can cause the coffee machine to dispense coffee in a specific way (e.g., black, with cream, but no sugar, etc.) in the morning in response to the detection of individual 310, based on the learning of corresponding usage patterns.Alternatively, a device such as a coffee machine with the Receiver 310 embedded in it can use audio to ask the person if they would like to take a specific action when it detects their presence and performs the corresponding identification.

[0046] Fig. Figure 3B shows an example of information that can be extracted from wireless signals for use in an embodiment of a method for determining a person's status, a person's status, or a combination thereof. In particular, it illustrates Fig. 3B for a single packet addressed to one recipient (e.g., recipient 304 of Fig. 3A) is received, CSI amplitude information 352, CSI phase information 354 and CSI received signal strength indicator (RSSI) information 356.

[0047] As above with reference to Fig. As noted in section 3A, wireless signals undergo multipath propagation between a transmitter and a receiver. If interference is introduced into the various propagation paths, the amplitude, phase, and received signal strength of these signals at the receiver can be affected. Accordingly, if a moving person traverses the various propagation paths, the receiver can detect this movement by observing changes in amplitude, phase, and received signal strength. Furthermore, these changes in amplitude, phase, and received signal strength within a given packet or across multiple packets can be used to determine characteristics of the person's movement. For example, a person's gait is unique to each individual due to various factors.Accordingly, changes in amplitude, phase, and received signal strength within a given packet or across multiple packets can be used to determine a specific person's gait during the training of a machine learning model and later for classification and thus identification of that person. This information can be combined with other information (e.g., the person's usage patterns and routines) to trigger additional actions by devices associated with the recipient.

[0048] Fig. Figure 3C shows a workflow diagram for an embodiment of a method according to the disclosure. Method 360, in the embodiment shown, can be implemented in a variety of ways, which are contained in or associated with a wireless receiver. Method 360 includes wireless signal acquisition (Block 362), data preprocessing (Block 364), an annotation process (Block 366), classifier training (Block 368), and inference (Block 370). These various operations will now be explained in more detail.

[0049] For wireless signal data collection (Block 362), a receiver can be used to capture wireless signals and gather information from them. For example, a Wi-Fi receiver can receive wireless signals in packets and collect CSI (e.g., amplitude and phase) from the received packets. Variations in amplitude and phase detected in the CSI within individual packets, as well as across multiple packets, can be used to track human movement within a location (e.g., a room in a home) where both the transmitter and receiver are located.

[0050] The transmitter can be, in various embodiments, a Wi-Fi router already present on site, although other types of transmitters (e.g., Bluetooth) are possible and are considered within the scope of this disclosure. The receiver, as mentioned above, can be embedded in a device or other equipment capable of using the information obtained from the CSI of the received packets. In various embodiments, the receiver can be connected to a laptop / tablet / phone to control the CSI collection and visually review this data during the collection phase.

[0051] CSI traces within packets can be used to explicitly capture time-domain information (e.g., variations in amplitude over time) as well as frequency-domain information. These traces can be particularly sensitive to human movement within the environment. Accordingly, amplitude, phase, and received signal strength indicator (RSSI) values ​​can be determined from each CSI trace to obtain different perspectives of the channel frequency response resulting from human movement.

[0052] Data preprocessing (Block 364) can perform various processing tasks to minimize the effects of noise in the received packets. As noted above, CSI can be particularly sensitive to human movement and, more generally, to the environment as a whole, since wireless signals propagate to the receiver via multiple paths and thus multiple reflections. For example, signals in a space can be reflected by furniture, walls, ceilings, floors, cabinets, and other inanimate objects. Furthermore, the presence of other signal sources in the area (e.g., another Wi-Fi access point operating on the same channel) can introduce additional interference or noise into the collected data. These factors can make the CSI signals extremely noisy.To reduce the influence of unwanted noise and improve human motion detection, the 360 ​​method can apply a series of preprocessing steps to clean the signal data. For example, raw CSI data contains null and pilot subcarriers, which are part of the Orthogonal Frequency Division Multiplexing (OFDM) stack to ensure less interference for users operating on multiple frequency channels. These subcarriers are removed to reduce dimensionality and redundancy, as they also carry no useful information. CSI segments with high human motion content can then be selected using the annotated labels during data acquisition.

[0053] One embodiment, described below with reference to Fig. As discussed in 4A, preprocessing is performed by using the CSI amplitude as the identification features. Another embodiment can use preprocessing that utilizes CSI after phase unpacking and phase cleanup, and is discussed further below with reference to Fig. 4B is discussed. Embodiments that use Doppler Frequency Shift (DFS) maps of raw, complex-valued CSI are also possible and are under consideration. In DFS embodiments, amplitude and phase values ​​can be cleaned according to the algorithms mentioned above, followed by performing a short-time Fourier transform on smaller overlapping CSI windows. The result of this workflow can be DFS maps that exhibit high human motion sensitivity. One embodiment uses statistical features as input to a machine learning algorithm. In the embodiment with statistical features, the preprocessed CSI amplitude or phase can be used to extract specific time- and frequency-domain information such as mean, variance, fast Fourier transforms (FFTs), and so on.It is noted that preprocessing using any combination of the features of these embodiments is also possible and will be considered.

[0054] The annotation process (Block 366) can be used for person status identification. In various embodiments, CSI samples corresponding to a person can be given an anonymous label to protect privacy, such as "Person A," "Child," or "Adult." The labels can remain the same for each person regardless of changes in the environment or the time of collection, to maintain consistency for the PID task. Depending on the number of people in the data collection and annotation phases, the type of classifier training can change to either binary or multi-class classification. In this phase, each time segment (e.g., 2 seconds of data) can be annotated with a class, such as "Person A," "Child," or "Adult." In an embodiment configured for use in a house, the total number of classes can be equal to the total number of person statuses in a house, or it could be one more (e.g.,(An unknown visitor or an empty house). These annotations can be collected by a user as an initial setup phase or collected in the background based on user interaction with a product, such as a coffee machine. For example, the coffee machine might monitor the residents' behavior for just seven days, and when someone dispenses a coffee, it takes the previous sequences of Wi-Fi data, labels them with a user profile, such as Person A, and uses this data for future reference.

[0055] Classifier training can be performed in various embodiments, the features of which can be combined. One embodiment can use a 2D convolutional neural network (CNN) based on standard ResNet18 residual connections and squeeze excitation (SE) blocks. Such a network can have and can assume a small number of parameters. The network has very few parameters (680K) and takes the spatiotemporal CSI amplitude as input, with the input being used to learn to predict the correct label corresponding to each element in provided datasets. The proposed neural network can, in one embodiment, include a 7x7 kernel-size input convolutional layer. In another embodiment, a different filter size can be used instead of a 7x7 filter, for example, a 51x51 filter.The large input convolutions in the input space can lead to the ingestion of more CSI information in the time and frequency domains, resulting in enhanced feature representation learning in the downstream convolution blocks. Moving to the deeper layers of the network, the residual identity blocks can preserve more information in subsequent deeper layers, while the single identity (SE) blocks can improve the network's representational performance by enabling dynamic channel-wise feature recalibration. Implementations can also use a ResNet18 model as a baseline to compare the performance of a network that can have a large number of parameters.

[0056] The training phase can end automatically when the models achieve high classification accuracy and consequently low training target loss on the validation sets. In addition to amplitude, phase information can also be provided as input to the neural network. Several separate approaches to using phase information can be implemented. For example, amplitude and phase can be concatenated as a long sequence and learned together by a single network (early fusion). In another example, two separate neural networks can be trained to learn amplitude and phase features separately and concatenated before the knowledge is distilled into classification layers (late fusion).

[0057] In another embodiment of the classifier training, performed according to Block 368, statistical features can be used for classification. In these embodiments, a comprehensive set of features can be extracted from Wi-Fi CSI data for use in machine learning models. Both time-domain and frequency-domain features can be extracted from the CSI data, which can be represented as a three-dimensional array with dimensions corresponding to samples, subcarriers, and time steps.

[0058] In various implementations, classifier training can involve performing principal component analysis (PCA): PCA is applied to transform the data, reducing its dimensionality while preserving essential information. This step involves reshaping each sample to align time steps and channels, and then applying PCA, resulting in a set of principal components that capture the most significant variations in the signal.

[0059] In some implementations, classifier training may include determining the rate of change of the RSSI, calculating the rate of change of the received signal strength indicator (RSSI) over a specified interval. This can provide insight into how the signal strength varies over time, which can aid in understanding the dynamics of the wireless channel.

[0060] Implementations of classifier training can also utilize time-domain features. Various statistical measures can be calculated across subcarriers, including mean, median, variance, standard deviation, skewness, kurtosis, root mean square (RMS), and zero-crossing rate. These features can help characterize the distribution and variability of signals in the time domain.

[0061] Implementations can also, or alternatively, use frequency domain features. A fast Fourier transform (FFT) can be applied to extract features such as spectral centroid, spectral bandwidth, spectral flatness, and peak frequency. These features capture essential properties of the signal's frequency content, such as its dominant frequency and spread.

[0062] Time-frequency plots can be used in some implementations. For example, short-time Fourier transform coefficients (STFT coefficients) can be calculated and analyzed, providing a combined view of how the frequency content of the signal evolves over time. Correlation analysis can also be used in some implementations. A subcarrier correlation can be calculated for each time point, providing insights into the relationships and dependencies between different parts of the spectrum.

[0063] The various methodologies used in classifier training can consolidate the different features discussed above into a two-dimensional feature matrix suitable for machine learning models. This transformation can include reshaping and bin averaging, which in turn can reduce the number of time steps required for the classification task. The feature matrix can then be smoothed and fed to classifiers such as SVM, XGBoost, MLP, Random Forest, Decision Tree, and so on.

[0064] Implementations of classifier training that use similar statistical features to those discussed above can also employ a sequence model for the classification task to capture the temporal pattern of human movement. Different sequence models can be used in different implementations.

[0065] In one embodiment, the sequence model can include a bidirectional gated recurrent unit (BiGRU) with an attention mechanism. The BiGRU layer can process input sequences in both forward and reverse directions, capturing dependencies over time steps. The attention mechanism can assign weights to different parts of the GRU output, focusing on sequence segments more relevant to the task at hand. The network can also include a dropout layer for regularization, which in turn can reduce the likelihood of overfitting. Finally, a fully connected layer can map the attention-weighted features to the desired output classes.This combination of bidirectional GRU and attention can be effective for tasks where understanding the context and importance of different parts of a sequence is a component. Statistical features can be fed into the BiGRU units at different time steps in various embodiments. In some embodiments, preprocessed CSI amplitude and / or CSI phase information can be fed into the BiGRU units at different time steps.

[0066] Implementations are also possible and are being considered in which bidirectional Long Short Term Memory (LSTM) with an attention mechanism is used in the classification process as an alternative to the GRU discussed above. Recurrent Neural Network (RNN) architectures can be used in various implementations.

[0067] In another embodiment, transformers are used to model the sequence. Several variations of the transformer can be used, including (a) transformers with static position coding, (b) transformers with learnable position coding, and (c) transformers with static position coding for the time domain and learnable position coding for the feature domain, and vice versa.

[0068] After classifier training, the trained models can be used for inference evaluation (Block 370) by keeping the parameters frozen. Inference can be performed either in real time or offline. Offline inference can follow steps that are the same as or similar to those discussed above, such as CSI amplitude processing and / or CSI phase processing, where the raw CSI samples are provided as input to the trained model. A procedure for performing online inference is described below with reference to Fig. 4D discussed.

[0069] Although in Fig. While not shown in Figure 3C, embodiments including an actuation phase are also possible and are being considered. In the actuation phase, a system that includes the receiver and executes the machine learning model can personalize the user experience of a device or other equipment. The actuation can involve the system performing various actions (e.g., dispensing coffee according to personal preferences) with minimal (if any) input from the user. It can also enable a child lock feature to prevent a child from operating a kitchen stove or washing machine.

[0070] Fig. Figure 4A shows a workflow diagram and signal characteristics for amplitude preprocessing of CSI for an embodiment of the disclosed method. As shown in Fig. As shown in Figure 4A, zero subcarriers can be removed from received packets, followed by noise removal using, for example, a Butterworth bandpass filter and Hanning smoothing. Regardless of the type of filter used, the filtering process can remove high- and very-low-frequency noise components corresponding to reflections from inanimate objects (e.g., furniture, cabinets, etc.). Other types of stationary and non-stationary signal processing filters, such as low-pass and wavelet-based methods, can also be implemented.

[0071] DC offset components can also be removed, followed by data normalization. To further reduce dimensionality, lossless frequency-based subsampling can be performed. Alternatively, PCA-based dimensionality reduction can be used instead of subsampling.

[0072] Procedure 400 includes CSI amplitude extraction (Block 402, illustrated graphically in 420), which involves determining the raw amplitude of wireless signals received in a packet. The procedure then proceeds with zero-subcarrier removal (Block 404, illustrated graphically in 404), since the zero subcarriers carry no information and are primarily used to shape the signal spectrum. The procedure then performs sample duration selection and shortening (Block 407), followed by filtering operations that include high-frequency component removal (Block 408) and DC component removal (Block 410). The filtering operations can be visualized as bandpass filtering in 424, smoothed CSI amplitude in 426, and DC-smoothed CSI amplitude in 428. Finally, the remaining data is normalized (Block 412, illustrated graphically in 430).The data are then undersampled to reduce dimensionality by performing lossless, frequency-based time-domain undersampling, as shown in 432, with these data being fed into the model for classifier training (Block 424).

[0073] Fig. Figure 4B shows a workflow diagram and signal characteristics for phase preprocessing of channel state information (CSI) for an embodiment of the disclosed method. In this example, the phase of the wireless signals of a packet is used to extract the data for classifier training.

[0074] Procedure 431 involves CSI phase extraction (block 434) from raw CSI phase data, illustrated graphically in 450. The zero subcarriers are removed and the phase data is unpacked (blocks 436 and 438, respectively, illustrated graphically in 452), followed by phase correction and smoothing (block 440, illustrated graphically in 454). Next, the CSI phase data is smoothed (illustrated graphically in 456) and DC components are subsequently removed (block 442, illustrated graphically in 458). Data normalization is then performed (block 444, illustrated graphically in 460). Finally, subsampling is performed (block 462) and the subsampled data is fed into the model for classifier training (block 446).

[0075] Fig. Figure 4C shows a neural network and a workflow diagram illustrating its operation for one embodiment of the disclosed method. In the embodiment shown, the neural network 480 includes an input convolution layer 471, which in this particular example has a core size of 7 x 7, although embodiments using filters of different sizes are possible and are under consideration. Generally speaking, the input convolutions can be large in various embodiments to allow the input of more CSI data in the time and / or frequency domains, which can lead to better representations in subsequent convolution blocks of the neural network 480.

[0076] The data output by the input convolution layer 471 can be normalized by the normalization layer 472. The normalized data output by the normalization layer is then provided to the maxpool layer 472, where it can be downsampled to reduce its spatial dimensions, which in turn can make the representation smaller and more manageable. The downsampled data can then be applied to the dropout layer 474 to prevent overfitting in the downstream layers. The operations performed in the dropout layer 474 can include the random ignoring of specific neurons and their respective connections. Neurons can be dropped at a specified rate, and this rate can be adjustable in various embodiments.

[0077] The data from dropout block 474 can then be applied to ResNet layer 475, where additional downsampling and convolution are performed. In this particular embodiment, the data output from this layer comprises 64 channels (and thus 64 feature maps) with a 3 x 3 kernel for the convolution operation. The data from ResNet layer 475 is then applied to the squeeze-excited layer 476 to further refine the classification by performing channel-wise feature recalibration. The data output from layer 476 then undergoes another dropout operation in dropout layer 477.

[0078] After the dropout operation from layer 477, the resulting data is applied to another ResNet layer 478, where further downsampling and convolution are performed using a convolution block comprising a 3 x 3 matrix. This downsampled data is then provided to another squeeze-excited layer 479, followed by another dropout operation in dropout layer 480. Afterward, the data from dropout layer 480 is applied to the averaging-pooling layer 471 to further reduce the spatial dimensions of the data. The reduced data is then provided to the linear softmax layer 481 for final classification. In the linear portion of layer 481, the output is provided as a combination of inputs, where a weight sum of inputs and a bias term are calculated.The softmax function serves as an activation function for a multi-class classification operation, which, in the context of this disclosure, may involve identifying one of a number of distinct persons. Since the output of the softmax function can be interpreted as a probability that the input data belongs to a particular class, the output, in the context of this disclosure, can provide a probability that the input data corresponds to a specific person.

[0079] Fig. Figure 4D is a workflow diagram illustrating an embodiment of a method for performing inference according to the disclosure. In particular, the method 489, as shown in the illustrated embodiment, can represent a real-time inference methodology that uses input data after training a machine learning model and thus the process of identifying a specific person. In this example, the CSI traces from real-time packets can be fed into the model as a continuous data stream. Window-based slicing can be performed to obtain smaller pieces of the time-series data and to carry out the subsequent preprocessing steps, which are similar to training data creation. After signal cleaning, the signal can be fed as input to the classifier to output a class prediction for each moving window.

[0080] Method 489 includes real-time CSI window processing (Block 484) and CSI amplitude extraction (Block 486). It is noted that, although Method 489 is discussed herein using CSI amplitude information, CSI phase information or other information may be used in other embodiments. Zero subcarrier removal (Block 488), high-frequency noise removal (Block 490), DC component removal (Block 492), and data normalization (Block 494) are also performed in the same or a similar manner as during the training process. As a result, data is produced that can then be provided for inference (Block 496) to perform the identification of a person as discussed above.

[0081] Fig. 4E generally presents a method 4000 for identifying a person's status according to the principles of this disclosure. In 4002, the method 4000, at a wireless receiver, collects channel state information from received packets sent by a wireless transmitter and performs preprocessing. For example, the computer system 202, at the wireless receiver 304, can collect the channel state information from received packets sent by the wireless transmitter 302.

[0082] In 4004, procedure 4000, using computer system 202, annotates the selected channel state information segments with a class indicating a status (e.g., child status or adult status) of the person, based on a gait and at least some of a variety of biometric features of the person extracted from the channel state information.

[0083] In 4006, the procedure 4000, using a machine learning model (such as machine learning model 210), identifies the person's status based on their gait and at least some of the person's numerous biometric characteristics. Machine learning model 210 can be trained using classifier training and training data that includes information from selected channel state information segments and gait and biometric information from individuals with different statuses. In some embodiments, performing classifier training on the machine learning model involves using a two-dimensional convolutional neural network to determine the person's status.In some embodiments, performing classifier training of the machine learning model includes: extracting a variety of time-domain features to determine the variability of wireless signals of the selected channel state information segments; and extracting a variety of frequency-domain features to determine the spectral bandwidth, spectral flatness, and peak frequency of the wireless signals of the selected channel state information segments, including subcarrier correlations. In some embodiments, performing the classifier training further includes using a sequence model with a bidirectional gated recurrent unit (BiGRU) containing an attention mechanism and a transformer.In some embodiments, performing the classifier training of the machine learning model involves using a sequence model to determine a temporal movement pattern of the person.

[0084] In procedure 4008, using computer system 202 and machine learning model 210, and based on the person's status, determines one or more settings of a machine located at a site that includes wireless receiver 304 and wireless transmitter 302. The one or more settings can correspond to the person's status (e.g., the one or more settings can be set by a user of the machine and can, without restriction, include locking a door of the machine in response to the person's status indicating that the person is a child).

[0085] In 4010, procedure 4000, using computer system 210 and based on one or more settings, controls the operation of the machine in response to the detection of the person's status.

[0086] Fig. Figure 5 shows a schematic diagram of an interaction between a computer-controlled machine 500 and a control system 502. The computer-controlled machine 500 includes an actuator 504 and a sensor 506. The actuator 504 can include one or more actuators, and the sensor 506 can include one or more sensors. The sensor 506 is configured to detect a state of the computer-controlled machine 500. The sensor 506 can be configured to encode the detected state into sensor signals 508 and transmit sensor signals 508 to the control system 502. Non-restrictive examples of the sensor 506 include wireless receivers, video, radar, LiDAR, ultrasonic, and motion sensors, as described above with reference to Fig. 1-2 described. In one embodiment, the sensor 506 is a wireless sensor configured to detect an environment near the computer-controlled machine 500. Embodiments in which a combination of different sensors is also possible and is being considered.

[0087] The Sensor 506 can also be a wireless signal receiver in various embodiments, configured to receive wireless signals from a transmitter (e.g., Wi-Fi). The computer-controlled machine can use the received wireless signals for personal identification in various embodiments, based on the detection of a person's movement in the vicinity and characteristics of that movement, which are used to train a machine learning model.

[0088] The control system 502 is configured to receive sensor signals 508 from the computer-controlled machine 500. As explained below, the control system 502 can further be configured to calculate actuator control commands 510 based on the sensor signals and to transmit actuator control commands 510 to the actuator 504 of the computer-controlled machine 500.

[0089] As in Fig. As shown in Figure 5, the control system 502 includes the receiver unit 512. The receiver unit 512 can be configured to receive sensor signals 508 from the sensor 506 and transform the sensor signals 508 into input signals x. In an alternative embodiment, the sensor signals 508 are received directly as input signals x without the receiver unit 512. Each input signal x can be a part of each sensor signal 508. The receiver unit 512 can be configured to process each sensor signal 508 to generate each input signal x. The input signal x can contain data corresponding to an image recorded by the sensor 506.

[0090] The control system 502 includes a classifier 514. The classifier 514 can be configured to classify the input signals x into one or more labels using a machine learning (ML) algorithm, such as a neural network described above. The classifier 514 is configured to be parameterized by parameters, such as those described above (e.g., parameter θ). The parameters θ can be stored in and provided by the non-volatile memory 516. The classifier 514 is configured to determine the output signals y from the input signals x. Each output signal y contains information that assigns one or more labels to each input signal x. The classifier 514 can send the output signals y to the conversion unit 518. The conversion unit 518 is configured to convert the output signals y into actuator control commands 510.The control system 502 is configured to transmit the actuator control commands 510 to the actuator 504, which is configured to actuate the computer-controlled machine 500 in response to the actuator control commands 510. In another embodiment, the actuator 504 is configured to actuate the computer-controlled machine 500 directly based on the output signals y.

[0091] Upon receiving the actuator control commands 510, the actuator 504 is configured to perform an action corresponding to the associated actuator control command 510. The actuator 504 may include control logic configured to transform the actuator control commands 510 into a second actuator control command, which is used to control the actuator 504. In one or more embodiments, the actuator control commands 510 can be used to control a display instead of, or in addition to, an actuator.

[0092] In another embodiment, the control system 502 includes the sensor 506 instead of or in addition to the computer-controlled machine 500, which includes the sensor 506. The control system 502 can also include the actuator 504 instead of or in addition to the computer-controlled machine 500, which includes the actuator 504.

[0093] As in Fig. As shown in Figure 5, the control system 502 also includes the processor 520 and the memory 522. The processor 520 can include one or more processors. The memory 522 can include one or more memory devices. The classifier 514 (e.g., machine learning algorithms such as those described above with respect to the pretrained classifier 306) of one or more embodiments can be implemented by the control system 502, which includes the non-volatile memory 516, the processor 520, and the memory 522.

[0094] The non-volatile memory 516 can include one or more persistent data storage devices such as a hard disk, optical drive, tape drive, non-volatile solid-state device, cloud storage, or any other device capable of permanently storing information. The processor 520 can include one or more devices selected from high-performance computing (HPC) systems, including high-performance cores, microprocessors, microcontrollers, digital signal processors, microcomputers, central processing units, field-programmable gate arrays, programmable logic devices, state machines, logic circuits, analog circuits, digital circuits, or any other devices that manipulate signals (analog or digital) based on computer-executable instructions located in memory 522.The memory 522 can include a single memory device or a number of memory devices, including but not limited to random access memory (RAM), volatile memory, non-volatile memory, static random access memory (SRAM), dynamic random access memory (DRAM), flash memory, cache memory, or any other device capable of storing information.

[0095] The 520 processor can be configured to read into the 522 memory and execute computer-executable instructions located in the 516 non-volatile memory, which embody one or more machine learning algorithms and / or methodologies of one or more implementations. The 516 non-volatile memory can contain one or more operating systems and applications. The 516 non-volatile memory can store compiled and / or interpreted computer programs created using a variety of programming languages ​​and / or technologies, including, without limitation and either alone or in combination, Java, C, C++, C#, Objective-C, Fortran, Pascal, JavaScript, Python, Perl, and PL / SQL.

[0096] When executed by the processor 520, the computer-executable instructions in the non-volatile memory 516 can cause the control system 502 to implement one or more of the ML algorithms and / or methodologies disclosed herein. The non-volatile memory 516 can also contain ML data (including data parameters) that support the functions, features, and processes of one or more embodiments described herein.

[0097] The program code embodying the algorithms and / or methodologies described herein may be distributed individually or collectively as a program product in a variety of different forms. The program code may be distributed using a computer-readable storage medium containing computer-readable program instructions to instruct a processor to execute aspects of one or more embodiments. Computer-readable storage media, which are inherently non-volatile, may include volatile and non-volatile, removable and non-removable physical media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data.Computer-readable storage media may also include RAM, ROM, erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other solid-state storage technology, portable compact disc read-only storage (CD-ROM) or other optical storage, magnetic cartridges, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and that can be read by a computer. Computer-readable program instructions may be downloaded to a computer, other type of programmable data processing device, or other device from a computer-readable storage medium, or to an external computer or external storage device over a network.

[0098] Computer-readable program instructions stored on a computer-readable medium can be used to instruct a computer, other types of programmable data processing devices, or other devices to operate in a specific manner, such that the instructions stored on the computer-readable medium produce a manufactured article containing instructions that implement the functions, actions, and / or operations specified in the flowcharts or diagrams. In certain alternative embodiments, the functions, actions, and / or operations specified in the flowcharts and diagrams can be reordered, processed serially, and / or processed simultaneously in accordance with one or more embodiments.Furthermore, each of the flowcharts and / or diagrams may contain more or fewer nodes or blocks than those illustrated in accordance with one or more embodiments.

[0099] The processes, procedures or algorithms can be embodied wholly or partially using suitable hardware components, such as application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), state machines, controllers or other hardware components or devices, or a combination of hardware, software and firmware components.

[0100] Fig. Figure 6 shows a schematic diagram of the control system 502, which is configured to control the vehicle 600, which can be at least a semi-autonomous vehicle or a semi-autonomous robot. The vehicle 600 includes the actuator 504 and the sensor 506. The sensor 506 can include one or more video sensors, cameras, radar sensors, ultrasonic sensors, wireless transmitters and / or receivers, LiDAR sensors, and / or position sensors (e.g., GPS). One or more of these specific sensors can be integrated into the vehicle 600. Alternatively, or in addition to one or more of the specific sensors identified above, the sensor 506 can include a software module configured to determine a state of the actuator 504 when executed.

[0101] The classifier 514 of the control system 502 of the vehicle 600 can be configured to detect objects near the vehicle 600 based on the input signals x. In such an embodiment, the output signal y can contain information characterizing the proximity of objects to the vehicle 600. The actuator control command 510 can be determined according to this information. The actuator control command 510 can be used to avoid collisions with the detected objects. In some embodiments, the classifier 514 can use wireless signals (e.g., Bluetooth signals) in the vehicle for PID purposes as discussed above. For example, the classifier 514 can use the wireless signals to identify a specific driver of the vehicle, thus enabling the control system 502 to set a seat position for that driver upon entry into the vehicle.

[0102] In embodiments where the vehicle 600 is at least partially autonomous, the actuator 504 can be embodied in a brake, drive system, motor, powertrain, or steering system of the vehicle 600. The actuator control commands 510 can be determined such that the actuator 504 is controlled to prevent the vehicle 600 from colliding with detected objects. Detected objects can also be classified according to what the classifier 514 considers most likely, such as pedestrians or trees. The actuator control commands 510 can be determined depending on the classification. In a scenario where an adversary attack may occur, the system described above can further be trained to better detect objects or to identify changes in lighting conditions or angles for a sensor or camera on the vehicle 600.

[0103] In other embodiments, where the vehicle 600 is at least a semi-autonomous robot, the vehicle 600 can be a mobile robot configured to perform one or more functions, such as flying, swimming, diving, and kicking. The mobile robot can be at least a semi-autonomous lawnmower or at least a semi-autonomous cleaning robot. In such embodiments, the actuator control command 510 can be configured to control a drive unit, steering unit, and / or brake unit of the mobile robot so that the mobile robot can avoid collisions with identified objects.

[0104] In another embodiment, the vehicle 600 is an at least semi-autonomous robot in the form of a garden robot. In such an embodiment, the vehicle 600 can use an optical sensor or a wireless receiver and / or transmitter as sensor 506 to determine the condition of plants in an environment near the vehicle 600. The actuator 504 can be a nozzle configured to spray chemicals. Depending on an identified species and / or condition of the plants, the actuator control command 510 can be determined to cause the actuator 504 to spray the plants with an appropriate amount of suitable chemicals.

[0105] The vehicle 600 can be at least a semi-autonomous robot in the form of a household appliance. Non-restrictive examples of household appliances include a washing machine, a stove, an oven, a microwave, a coffee maker, or a dishwasher. In such a vehicle 600, the sensor 506 can be an optical sensor or a wireless receiver and / or transmitter configured to detect the state of an object to be processed by the household appliance. For example, if the household appliance is a washing machine, the sensor 506 can detect the state of the laundry inside the washing machine. The actuator control command 510 can then be determined based on the detected state of the laundry.In another example, if the household appliance is an oven, person status identification can be performed by the control system based on wireless signals received by the sensors 506 (in this example, a wireless receiver), with the oven configured to take one or more safety actions (e.g., looking at the oven door) according to one or more settings and based on the person's status.

[0106] Fig. Figure 7 shows a schematic diagram of the control system 502, which is configured to control a system 700 (e.g., a manufacturing machine), such as a punch cutter, a cutter, or a pistol drill, or a manufacturing system 702, such as a part of a production line. The control system 502 can be configured to control the actuator 504, which is configured to control the system 700 (e.g., the manufacturing machine).

[0107] The sensor 506 of the system 700 (e.g., the manufacturing machine) can be an optical sensor (such as those described above) configured to detect one or more properties of the manufactured product 704. The classifier 514 can be configured to determine a condition of the manufactured product 704 from one or more of the detected properties. The actuator 504 can be configured to control the system 700 (e.g., the manufacturing machine) based on the determined condition of the manufactured product 704 for a subsequent manufacturing step or for binning the manufactured product 704 (e.g., discarding, sorting, marking, trimming, or repairing) if the manufactured product 704 has a detected defect. The actuator 504 can be configured to transfer functions of the system 700 (e.g., the manufacturing machine) to the subsequently manufactured product 706 of the system 700 (e.g.,to control the manufacturing machine) depending on the specific state of the manufactured product 704.

[0108] In some embodiments, the sensor 506 can be coupled to receive and process wireless signals (e.g., Wi-Fi signals), while the control system 502 can use the information obtained for a PID as discussed elsewhere herein. For example, using wireless signals detected by the sensor 506, the control system 502 can identify the status of a person who regularly works with the system 700. Based on the status of that person, the control system 502 can make adjustments to the system 700 (e.g., to the speed of a conveyor belt).

[0109] Fig. Figure 8 shows a schematic diagram of the control system 502, which is configured to control a power tool 800, such as an electric drill or screwdriver, which has at least a semi-autonomous mode. The control system 502 can be configured to control the actuator 504, which is configured to control the power tool 800.

[0110] The sensor 506 of the power tool 800 can be an optical sensor configured to detect one or more properties of the work surface 802 and / or the fastener 804 driven into the work surface 802. The classifier 514 can be configured to determine a state of the work surface 802 and / or the fastener 804 relative to the work surface 802 from one or more of the detected properties. The state can be that the fastener 804 is flush with the work surface 802. Alternatively, the state can be the hardness of the work surface 802. The actuator 504 can be configured to control the power tool 800 such that the drive function of the power tool 800 is adjusted depending on the determined state of the fastener 804 relative to the work surface 802 or on one or more detected properties of the work surface 802.For example, the actuator 504 can interrupt the drive function when the fastening element 804 is flush with the work surface 802. As another non-restrictive example, the actuator 504 can apply additional or less torque depending on the hardness of the work surface 802.

[0111] In some embodiments, the sensor 506 can be or include a wireless receiver for detecting the presence of wireless signals (e.g., Wi-Fi, Bluetooth signals). As discussed elsewhere herein, the control system 502 can use the wireless signals detected by the sensor 506 to determine the status of a specific person. Based on the status of that person, the control system 502 and the actuator 504 can make one or more adjustments to the operation of the power tool.

[0112] Fig. Figure 9 shows a schematic diagram of the control system 502, which is configured to control an automated personal assistant 900. The control system 502 can be configured to control the actuator 504, which is configured to control the automated personal assistant 900. The automated personal assistant 900 can be configured to control a household appliance, such as a washing machine, stove, oven, microwave, or dishwasher.

[0113] The sensor 506 can be an optical sensor and / or an audio sensor. The optical sensor can be configured to receive video images of gestures 904 from the user 902. The audio sensor can be configured to receive a voice command from the user 902. The sensor 506 can also, or alternatively, be configured to receive and process wireless signals (e.g., Wi-Fi), while the control system 502 can be configured to execute a PID as discussed elsewhere in this disclosure.

[0114] The control system 502 of the automated personal assistant 900 can be configured to determine actuator control commands 510, which are configured to control the system 502. The control system 502 can be configured to determine actuator control commands 510 according to the sensor signals 508 of the sensor 506. The automated personal assistant 900 is configured to transmit the sensor signals 508 to the control system 502. The classifier 514 of the control system 502 can be configured to execute a gesture recognition algorithm to identify the gesture 904 performed by the user 902, determine the actuator control commands 510, and transmit the actuator control commands 510 to the actuator 504. The classifier 514 can be configured to retrieve information from non-volatile memory in response to gesture 904 and output the retrieved information in a form suitable for receipt by user 902.In some embodiments, the classifier 514 can be configured to identify a specific person using the information received from the sensor 506, causing the control system 502 to stop operating the automated personal assistant 900 based on the identification.

[0115] Fig. Figure 10 shows a schematic diagram of the control system 502, which is configured to control the monitoring system 1000. The monitoring system 1000 can be configured to physically control access through the door 1002. The sensor 506 can be configured to detect a scene relevant to the decision of whether to grant access. The sensor 506 can be an optical sensor configured to generate and transmit image and / or video data. Such data can be used by the control system 502 to detect a person's face. In some embodiments, the sensor 506 may include (or alternatively include) circuitry for receiving and processing wireless signals such as Wi-Fi or Bluetooth signals.

[0116] The classifier 514 of the control system 502 of the monitoring system 1000 can be configured to interpret the image and / or video and / or wireless data by matching it with the identities of known persons stored in the non-volatile memory 516, thereby determining a person's identity. The classifier 514 can be configured to generate an actuator control command 510 in response to the interpretation of the image and / or video data. The control system 502 is configured to transmit the actuator control command 510 to the actuator 504. In this embodiment, the actuator 504 can be configured to lock or unlock the door 1002 in response to the actuator control command 510. In other embodiments, non-physical, logical access control is also possible.In some embodiments, the classifier 514 can be configured to generate an actuator control command based on the identification of a specific person based on wireless signals received by a wireless receiver in the sensor 506. For example, the classifier 514 can generate a command to cause the actuator 504 to set a temperature based on the status of the person identified based on wireless signals.

[0117] The monitoring system 1000 can also be a surveillance system. In such an embodiment, the sensor 506 can be an optical sensor or a wireless receiver and / or a wireless transmitter configured to detect a scene being monitored, and the control system 502 is configured to control the display 1004. The classifier 514 is configured to determine a scene classification, for example, whether the scene detected by the sensor 506 is suspicious. The control system 502 is configured to transmit an actuator control command 510 to the display 1004 in response to the classification. The display 1004 can be configured to adjust the displayed content in response to the actuator control command 510. For example, the display 1004 can highlight an object that the classifier 514 considers suspicious.Using an embodiment of the disclosed system, the surveillance system can predict objects that will appear at specific times in the future.

[0118] Fig.Figure 11 shows a schematic diagram of the control system 502, which is configured to control an imaging system 1100, such as an MRI, X-ray imaging device, or ultrasound device. The sensor 506 can, for example, be an imaging sensor. The classifier 514 can be configured to determine a classification of all or part of the acquired image. The classifier 514 can be configured to determine or select an actuator control command 510 in response to the classification received from the trained neural network. For example, the classifier 514 can interpret a region of an acquired image as potentially anomalous. In this case, the actuator control command 510 can be determined or selected to cause the display 1102 to display the imaging and highlight the potentially anomalous region.

[0119] In various embodiments, the sensor 506 can also include circuits for receiving and processing wireless signals. The classifier 514 can use the wireless signals to identify a person's status. Based on this identification, the classifier 514 can initiate the generation of one or more commands to control the imaging system 1100.

[0120] In some embodiments, a method for determining a person's status using wireless signals includes collecting, at a wireless receiver, channel state information from received packets transmitted by a wireless transmitter; annotating, using the computer system, selected channel state information segments with a class indicating a person's status based on a gait and at least some of a variety of biometric features of the person extracted from the channel state information;Identifying, using a machine learning model, the status of the person based on their gait and at least some of the person's multiple biometric characteristics, wherein the machine learning model is trained using classifier training and training data comprising information from selected channel state information segments and gait and biometric information from persons with different statuses; determining, using the computer system and the machine learning model and based on the person's status, one or more settings of a machine located in a space comprising the wireless receiver and the wireless transmitter, wherein the one or more settings correspond to the person's status;and control, by the computer system and using one or more settings, of the operation of the machine in response to the detection of the person's status;

[0121] In some embodiments, preprocessing the channel state information involves using amplitude information from the received packets to determine the person's gait and some of their many biometric characteristics. In some embodiments, preprocessing the channel state information involves using phase information from the received packets to determine the person's gait and some of their many biometric characteristics. In some embodiments, performing classifier training on the machine learning model involves using a two-dimensional convolutional neural network to determine the person's status.In some embodiments, performing classifier training of the machine learning model includes: extracting a variety of time-domain features to determine the variability of wireless signals of the selected channel state information segments; and extracting a variety of frequency-domain features to determine the spectral bandwidth, spectral flatness, and peak frequency of the wireless signals of the selected channel state information segments, including subcarrier correlations. In some embodiments, performing the classifier training further includes using a sequence model with a bidirectional gated recurrent unit (BiGRU) containing an attention mechanism and a transformer.In some embodiments, performing the classifier training of the machine learning model involves using a sequence model to determine a temporal movement pattern of the person. In some embodiments, the machine is a household appliance. In some embodiments, the person's status indicates whether the person is a child or an adult.

[0122] In some embodiments, a system for determining a person's status using wireless signals comprises: a wireless receiver configured to receive packets transmitted by a wireless transmitter and further configured to collect channel state information from the packets; and a computer system associated with the wireless receiver, wherein the computer system is configured to: annotate the selected channel state information segments with a class indicating a status of the person based on a gait and at least some of a variety of biometric features of the person extracted from the channel state information;to identify the status of a person using a machine learning model based on their gait and at least some of the person's multiple biometric characteristics, wherein the machine learning model is trained using classifier training and training data comprising information from the selected channel state information segments and gait and biometric information from persons with different statuses; to determine, using the machine learning model and based on the person's status, one or more settings of a machine located in a space comprising the wireless receiver and the wireless transmitter, wherein the one or more settings correspond to the person's status; and to control the operation of the machine using the one or more settings in response to the detection of the person's status.

[0123] In some embodiments, preprocessing the channel state information involves using amplitude information from the received packets to determine the person's gait and some of their many biometric characteristics. In some embodiments, preprocessing the channel state information involves using phase information from the received packets to determine the person's gait and some of their many biometric characteristics. In some embodiments, performing classifier training on the machine learning model involves using a two-dimensional convolutional neural network to determine the person's status.In some embodiments, performing classifier training of the machine learning model includes: extracting a variety of time-domain features to determine the variability of wireless signals of the selected channel state information segments; and extracting a variety of frequency-domain features to determine the spectral bandwidth, spectral flatness, and peak frequency of the wireless signals of the selected channel state information segments, including subcarrier correlations. In some embodiments, performing the classifier training further includes using a sequence model with a bidirectional gated recurrent unit (BiGRU) containing an attention mechanism and a transformer.In some embodiments, performing the classifier training of the machine learning model involves using a sequence model to determine a temporal movement pattern of the person. In some embodiments, the machine is a household appliance. In some embodiments, the person's status indicates whether the person is a child or an adult.

[0124] In some embodiments, a non-volatile, computer-readable medium stores instructions which, when executed by a computer system, cause the computer system to perform operations that include: annotating the selected channel state information segments with a class indicating a person's status based on their gait and at least some of a variety of the person's biometric features extracted from the channel state information; and identifying, using a machine learning model, the person's status based on their gait and at least some of the person's biometric features, the status indicating whether the person is a child or an adult.Determine, using the machine learning model and based on the person's status, one or more settings of a machine located at a site containing the wireless receiver, wherein the one or more settings correspond to the person's status; and control, using the one or more settings, the operation of the machine in response to the detection of the person's status.

[0125] In some embodiments, the machine includes a household appliance.

[0126] While exemplary embodiments are described above, it is not intended that these embodiments describe all possible forms included in the claims. The terms used in the description are descriptive and not limiting, and it is understood that various modifications can be made without departing from the basic idea and scope of the disclosure. As previously described, the features of different embodiments can be combined to form further embodiments of the invention that may not be explicitly described or illustrated.While various embodiments may be described as offering advantages or being preferred over other embodiments or implementations of the prior art with respect to one or more desired properties, the average person skilled in the art recognizes that one or more features or properties may be compromised to achieve desired overall system attributes, which depend on the specific application and implementation. These attributes may include, but are not limited to, cost, strength, durability, life-cycle costs, marketability, appearance, packaging, size, ease of maintenance, weight, manufacturability, ease of assembly, etc.Insofar as embodiments are described as less desirable than other embodiments or implementations of the prior art with respect to one or more properties, these embodiments are therefore not outside the scope of the disclosure and may be desirable for certain applications.

Claims

[1] Method for determining a person’s status using wireless signals, the method comprising: Collecting, at a wireless receiver, channel state information from received packets sent by a wireless transmitter; Annotating, using a computer system, selected channel state information segments with a class indicating a status of the person, based on a gait and at least some of a variety of biometric characteristics of the person extracted from the channel state information; Identify, using a machine learning model, the status of the person based on their gait and at least some of the multitude of biometric features of the person, wherein the machine learning model is trained using classifier training and training data that includes information from the selected channel state information segments and gait and biometric information from persons with different statuses; Determine, using the computer system and the machine learning model and based on the person's status, one or more settings of a machine located at a site that includes the wireless receiver and the wireless transmitter, wherein the one or more settings correspond to the person's status; and Control, through the computer system and using one or more settings, of the operation of the machine in response to the detection of the person's status. [2] Method according to claim 1, further comprising preprocessing the channel state information using amplitude information from the received packets to determine the gait and some of the multitude of biometric features of the person. [3] Method according to claim 2, wherein the preprocessing of the channel state information comprises using phase information from the received packets to determine the gait and some of the multitude of biometric features of the person. [4] Method according to claim 1, wherein performing classifier training of the machine learning model comprises using a two-dimensional neural convolutional network to determine the status of the person. [5] Method according to claim 1, wherein performing classifier training of the machine learning model comprises: Extracting a variety of time-domain features to determine variability of wireless signals in the selected channel state information segments; and Extracting a variety of frequency domain characteristics to determine a spectral bandwidth, spectral flatness, and peak frequency of the wireless signals of the selected channel state information segments, including subcarrier correlations. [6] Method according to claim 5, wherein performing the classifier training further comprises using a sequence model with a bidirectionally gated recurrent unit (BiGRU) with an attention mechanism and a transformer. [7] Method according to claim 1, wherein performing the classifier training of the machine learning model comprises using a sequence model to determine a temporal movement pattern of the person. [8] Method according to claim 1, wherein the machine is a household appliance. [9] Method according to claim 1, wherein the status of the person indicates whether the person is a child or an adult. [10] System for determining a person's status using wireless signals, the system comprising: a wireless receiver configured to receive packets sent by a wireless transmitter and further configured to collect channel state information from the packets; and a computer system assigned to the wireless receiver, wherein the computer system is configured to: Annotating the selected channel state information segments with a class that indicates a person's status based on a gait and at least some of a variety of the person's biometric characteristics extracted from the channel state information; Identify, using a machine learning model, the status of the person based on their gait and at least some of the multitude of biometric features of the person, wherein the machine learning model is trained using classifier training and training data that includes information from the selected channel state information segments and gait and biometric information from persons with different statuses; Determine, using the machine learning model and based on the person's status, one or more settings of a machine located at a site that includes the wireless receiver and the wireless transmitter, wherein the one or more settings correspond to the person's status; and Controlling, using one or more settings, the operation of the machine in response to the detection of the person's status. [11] System according to claim 10, wherein the computer system is further configured to preprocess the channel state information using amplitude information from the received packets to determine the gait and some of the multitude of biometric features of the person. [12] System according to claim 11, wherein the preprocessing of the channel state information comprises using phase information from the received packets to determine the gait and some of the multitude of biometric features of the person. [13] System according to claim 10, wherein performing classifier training of the machine learning model comprises using a two-dimensional neural convolutional network to determine the status of the person. [14] System according to claim 10, wherein performing classifier training of the machine learning model comprises: Extracting a variety of time-domain features to determine the variability of wireless signals in the selected channel state information segments; and Extracting a variety of frequency domain characteristics to determine a spectral bandwidth, spectral flatness, and peak frequency of the wireless signals of the selected channel state information segments, including subcarrier correlations. [15] System according to claim 14, wherein performing the classifier training further comprises using a sequence model with a bidirectionally gated recurrent unit (BiGRU) with an attention mechanism and a transformer. [16] System according to claim 10, wherein performing the classifier training of the machine learning model comprises using a sequence model to determine a temporal movement pattern of the person. [17] System according to claim 10, wherein the machine is a household appliance. [18] System according to claim 10, wherein the status of the person indicates whether the person is a child or an adult. [19] Non-volatile, computer-readable medium that stores instructions which, when executed by a computer system, cause the computer system to perform operations that include: Annotating the selected channel state information segments with a class that indicates a person's status based on a gait and at least some of a variety of the person's biometric characteristics extracted from the channel state information; Identify, using a machine learning model, the status of the person based on their gait and at least some of the person's multiple biometric characteristics, where the status indicates whether the person is a child or an adult; Determine, using the machine learning model and based on the person's status, one or more settings of a machine located at a site containing the wireless receiver, wherein the one or more settings correspond to the person's status; and Controlling, using one or more settings, the operation of the machine in response to the detection of the person's status. [20] Computer-readable medium according to claim 19, wherein the machine includes a household appliance.

Images