Technologies for tracking objects within defined areas

Abstract

To provide technologies for tracking objects within various defined areas.SOLUTION: The technologies are for tracking various objects (e.g., mammals, animals, humans, pets) within various defined areas (e.g., rooms, apartments, residences, vehicles, tents) to determine whether those objects satisfy or do not satisfy various criteria, signatures or thresholds, which may relate to health, safety or security of those objects within those defined areas. Tracking the objects in the defined areas may be enabled via various radars (e.g., time-of-flight radars, Doppler radars) positioned within the defined areas. Some of such radars may operate in a Ku-band between about 12 GHz and about 18 GHz inclusive, a K-band between about 18 GHz and about 27 GHz inclusive, or a Ka-band between about 26.5 GHz and about 40 GHz inclusive, where each of the bands tracks those objects within the defined areas.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] (CROSS-REFERENCE TO RELATED PATENT APPLICATIONS) This patent application claims the benefit of priority to U.S. Provisional Patent Application No. 63 / 153,795, filed February 25, 2021, which is incorporated herein by reference for all purposes.

[0002] This patent application claims the benefit of priority to U.S. Provisional Patent Application No. 63 / 162,476, filed March 17, 2021, which is incorporated herein by reference for all purposes.

[0003] (Technical field) The present disclosure relates to tracking objects within a defined area. [Background technology]

[0004] A first person (e.g., caregiver, doctor, family member, social worker, home caregiver) may wish to track a second person (e.g., care recipient, patient) within a defined area (e.g., room, apartment) to ensure that the second person is safe, healthy, or secure within the defined area. However, doing so may be technically problematic for a variety of reasons. For example, the second person may want to maintain privacy regarding such tracking or ensure that such tracking is secure. Similarly, whatever technology the first person decides to use for such tracking (e.g., video camera, proximity sensor) may have various technical shortcomings (e.g., coverage gaps, insufficient accuracy). Summary of the Invention

[0005] The present disclosure enables various techniques for tracking various objects (e.g., animals, humans, pets) within various defined areas (e.g., rooms, apartments, homes, offices, tents, barracks, vehicles, aircraft, spacecraft, clinics, field clinics, hospitals, and field hospitals) to determine whether these objects meet or do not meet various criteria, signatures, or thresholds that may relate to the health, safety, or security of these objects within these defined areas or the environments in which they reside. These techniques can track objects within the defined areas through various radars (e.g., time-of-flight radar, Doppler radar, etc.) positioned within the defined areas. For example, some such radars can operate in the Ku band, which includes values between about 12 GHz and about 18 GHz, the K band, which includes values between about 18 GHz and about 27 GHz, or the Ka band, which includes values between about 26.5 GHz and about 40 GHz, each of which has been unexpectedly found to be technologically useful for tracking these objects within the defined areas.

[0006] One embodiment can include a method including the steps of providing a device to a user, the device including a processor and a time-of-flight radar, the processor coupled to the time-of-flight radar, the time-of-flight radar configured to operate in a K-band; and instructing the user to position the device within a defined area having live objects therein, activate the time-of-flight radar to operate in the K-band within the defined area, cause the time-of-flight radar operating in the K-band within the defined area to track live objects within the defined area, generate a set of data based on tracking the live objects within the defined area, transmit the set of data to the processor, and cause the processor to determine whether an object is experiencing an event within the defined area based on the set of data, and take action responsive to the event determined to be occurring within the defined area.

[0007] One embodiment may include a method that may include receiving, by a processor, a set of data from a time-of-flight radar operating in K-band within a defined area having a living object therein, wherein the time-of-flight radar generates the set of data based on operating in K-band within the defined area and tracks the living object within the defined area; determining, by the processor, based on the set of data, whether the object is experiencing an event within the defined area; and taking action responsive to the event determined to be occurring within the defined area by the processor.

[0008] One embodiment may include a system comprising a device including a processor and a time-of-flight radar, the processor coupled to the time-of-flight radar, the time-of-flight radar configured to operate in K-band, the device configured to be positioned within a defined area having living objects therein, the time-of-flight radar operating in K-band within the defined area tracking the living objects within the defined area, generating a set of data based on tracking the living objects within the defined area, transmitting the set of data to the processor, the processor determining based on the set of data whether the objects are experiencing an event within the defined area, and taking action in response to the event determined to be occurring within the defined area.

[0009] One embodiment can include a method including the steps of providing a device to a user, the device including a processor and a radar, the processor coupled to the radar; positioning the device within a defined area having an object therein; and instructing the user to activate the radar to operate within the defined area, such that the radar operating within the defined area tracks the object within the defined area, generate a set of data based on the tracking of the object within the defined area, transmit the set of data to the processor, such that the processor determines whether to perform an action based on the set of data, and take the action based on the set of data.

[0010] One embodiment may include a method including the steps of receiving, by a processor, a set of data from a radar operating within a defined area having an object living therein, generating the set of data based on the radar operating within the defined area and tracking the object within the defined area; determining, by the processor, whether an action should be taken based on the set of data; and taking, by the processor, the action based on the set of data.

[0011] One embodiment may include a system comprising a device including a processor and a radar, the processor coupled to the radar, the device configured to be positioned within a defined area having objects living therein, the radar within the defined area tracking the objects within the defined area, generating a set of data based on the tracking of the objects within the defined area, transmitting the set of data to the processor, the processor determining whether to take an action based on the set of data, and taking the action based on the set of data.

[0012] One embodiment may include a method including the steps of: positioning a device within a defined area having living objects therein, the device including a processor and a radar, the processor coupled to the radar; activating the radar to operate within the defined area, the radar operating within the defined area tracking objects within the defined area, generating a set of data based on the tracking of the objects within the defined area, transmitting the set of data to the processor, the processor determining whether to perform an action based on the set of data, and taking the action based on the set of data.

[0013] One embodiment may include a device comprising a vehicle including a processor, a radar, and an area, wherein the processor is coupled to the radar, the area configured to include a driver or passenger, the processor is programmed to activate the radar to track the driver or passenger within the area, generate a set of data based on the tracking of the driver or passenger within the area, send the set of data to the processor, and the processor is programmed to determine whether to take an action based on the set of data, and take the action based on the set of data. [Brief explanation of the drawings]

[0014] [Figure 1] FIG. 1 illustrates a top view of one embodiment of an area including a device including a radar according to the present disclosure. [Figure 2] FIG. 2 is a side view of FIG. 1 according to the present disclosure. [Figure 3] FIG. 1 illustrates an embodiment of a device including a radar according to the present disclosure. [Figure 4] FIG. 1 illustrates a set of embodiments of a set of devices each including a radar or sensor according to the present disclosure. [Figure 5] FIG. 1 is a logic diagram of one embodiment of a device including a radar according to the present disclosure. [Figure 6] FIG. 4 is an internal schematic diagram of FIG. 3 according to the present disclosure. [Figure 7] FIG. 1 is a logic diagram of a radar according to the present disclosure. [Figure 8] 6 shows a photograph of the internal cavity of FIG. 3 according to the present disclosure. [Figure 9] 1 illustrates a set of embodiments of a set of form factors embodying a radar according to the present disclosure. [Figure 10] FIG. 1 illustrates one embodiment of a circuit board with a set of antennas for a radar according to the present disclosure. [Figure 11] FIG. 4 illustrates one embodiment of coverage ranges for the devices of FIG. 3 according to the present disclosure. [Figure 12]8 illustrates one embodiment of a set of microphones for the device of FIG. 7 according to the present disclosure. [Figure 13] 8 illustrates one embodiment of a microphone for the device of FIG. 7 according to the present disclosure. [Figure 14] 8A-8C illustrate one embodiment of raw readings from the device of FIGS. 1-7 and a virtual skeleton formed by the device of FIGS. 1-7 from the raw readings according to the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0015] The present disclosure enables various techniques for tracking various objects (e.g., mammals, animals, humans, pets) within various defined areas (e.g., rooms, apartments, homes, vehicles, tents) and determining whether these objects meet or do not meet various criteria, signatures, or thresholds that may relate to the health, safety, or security of the objects within these defined areas. These techniques can track objects within the defined areas through various radars (e.g., time-of-flight radar, Doppler radar, etc.) positioned within the defined areas. For example, some such radars can operate in the Ku band, which includes values between about 12 GHz and about 18 GHz, the K band, which includes values between about 18 GHz and about 27 GHz, or the Ka band, which includes values between about 26.5 GHz and about 40 GHz, each of which has been unexpectedly found to be technologically useful for tracking these objects within the defined areas, as described further below.

[0016] The present disclosure will now be more fully described with reference to all of the accompanying drawings, in which several embodiments of the present disclosure are shown. However, the present disclosure may be embodied in many different forms and should not be construed as necessarily limited to the various embodiments disclosed herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the various concepts of the present disclosure to those skilled in the art. It should be noted that the same or similar numbering may refer to the same or similar elements throughout.

[0017] As used herein, various terms can mean direct or indirect, complete or partial, temporary or permanent, action or omission. For example, when an element is referred to as being "on," "connected," or "coupled" to another element, the element may be directly connected or coupled to the other element, or there may be intervening elements, including indirect or direct variations. In contrast, when an element is referred to as being "directly connected" or "directly coupled" to another element, there are no intervening elements present.

[0018] As used herein, the term "or" is intended to mean an inclusive "or," rather than an exclusive "or." That is, unless otherwise specified or clear from the context, "X utilizes A or B" is intended to mean any of the natural inclusive sequences. That is, if X utilizes A, if X utilizes B, or if X utilizes both A and B, "X utilizes A or B" is satisfied in any of the aforementioned cases. For example, "X includes A or B" means "X can include A," "X can include B," or "X can include A and B," unless otherwise specified or clear from the context.

[0019] As used herein, the singular terms "a," "an," and "the" are each intended to include intermediate integer or decimal forms (e.g., 0.0, 0.00, 0.000) and also include plural forms (e.g., 2, 3, 4, 5, 6, 7, 8, 9, 10, 100, hundreds, thousands, millions), unless the context clearly indicates otherwise. Similarly, the singular terms "a," "an," and "the" are each intended to mean "one or more," and the phrase "one or more" may nevertheless be used herein.

[0020] As used herein, the terms "comprises," "includes," or "comprising" each specify the presence of stated features, integers, steps, operations, elements, or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or groups thereof.

[0021] As used herein, when this disclosure states that something is "based on" something else, such statement refers to a basis that may also be based on one or more other things. In other words, unless expressly indicated otherwise, "based on" as used herein inclusively means "based at least in part on" or "based at least in part on."

[0022] As used herein, terms such as "next," "then," or other similar forms are not intended to limit the order of steps. Rather, these terms are merely used to guide the reader through this disclosure. While process flow diagrams may describe some operations as sequential processes, many of these operations may be performed in parallel or concurrently. Additionally, the order of operations may be rearranged.

[0023] As used herein, the terms "response" or "reaction" are intended to include machine-sourced actions or inactions, such as input (e.g., local, remote), or user-sourced actions or inactions, such as input (e.g., via a user input device).

[0024] As used herein, the terms "about" or "substantially" refer to a ±10% variation from the nominal value / term.

[0025] It should be noted that although various terms such as "first term," "second term," and "third term" may be used herein to describe various elements, components, regions, layers, or sections, these elements, components, regions, layers, or sections should not necessarily be limited by such terms. Rather, these terms are used to distinguish one element, component, region, layer, or section from another element, component, region, layer, or section. Thus, a first element, component, region, layer, or section described below could be referred to as a second element, component, region, layer, or section without departing from this disclosure.

[0026] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. These terms, as defined in commonly used dictionaries, should be interpreted as having a meaning consistent with the meaning in the context of the relevant art, and should not be interpreted in an idealized or overly formal sense unless expressly defined as such in this specification.

[0027] Features or functions described with respect to particular embodiments can be combined and subcombined in or with various other embodiments. Also, different aspects, components, or elements of the embodiments as disclosed herein can be similarly combined and subcombined. Furthermore, some embodiments, individually or collectively, may be components of larger systems, and other procedures may take precedence or otherwise modify their application. Furthermore, many steps may be required before, after, or concurrently with the embodiments as disclosed herein. It should be noted that any or all methods or processes as disclosed herein can be performed, at least in part, through at least one entity or party, in any manner.

[0028] All issued patents, published patent applications, and non-patent publications mentioned or cited in this disclosure are hereby incorporated by reference in their entirety for all purposes to the same extent as if each individual issued patent, published patent application, or non-patent publication was specifically and individually indicated to be incorporated by reference. For clarity, all incorporations by reference specifically include these incorporated publications for all purposes of this disclosure as if these specific publications were copy-pasted into this specification and were originally included in this disclosure. Accordingly, any reference to anything disclosed herein includes all subject matter incorporated by reference, as explained above. However, if a disclosure incorporated by reference into this disclosure conflicts in part or in whole with this disclosure, the present disclosure will control to the extent of the conflict or the broader disclosure or broader definition of a term. If such disclosures conflict in part or in whole with each other, the later disclosure will control to the extent of the conflict.

[0029] Figure 1 is a top view of one embodiment of an area including a device including a radar according to the present disclosure. Figure 2 is a side view of Figure 1 according to the present disclosure. Figure 7 is a logic diagram of a radar according to the present disclosure. Figure 9 shows an embodiment of a set of form factors embodying a radar according to the present disclosure. In particular, area 100 includes device 102, object 104, stand 106, sofa 108, table 110, chair 112, oven 114, refrigerator 116, bathroom 118, toilet 120, bathtub 122, sink 124, outlet 126, wall 128, field of view 130, wall 132, door 134, and door 138.

[0030] Objects 104 may include mammals, animals, humans, pets, or any other suitable objects that may reside or be present in area 100, whether male or female. Mammals may include animals, humans, pets, or any other suitable animals. Animals may include zoo animals, humans, pets, or any other suitable animals. Humans may be infants, toddlers, preschoolers, school-age children, teenagers, young adults, adults, or seniors. Pets may include dogs, cats, rabbits, birds, or any other suitable pets. Note that objects may not live in area 100 but may be present in area 100. For example, this may apply to visitors, workers, maintenance personnel, cleaning personnel, medical personnel, emergency personnel, or other objects (e.g., mammals, animals, humans, pets) that may be animate or inanimate, living or not living within area 100, or that may be present or mobile within area 100.

[0031] The area 100 is embodied as a residence (e.g., a studio apartment) for the object 104. As shown in FIG. 1 (top view), the area 100 is defined by walls 132, doors 134, and windows 136, resulting in a rectangular shape. However, it should be noted that the area 100 may be defined by the walls 132, doors 134, or windows 136 in another suitable shape (e.g., square, oval, polygon, open-ended, closed-ended, teardrop, or cornerless). The walls 132 include a stud (e.g., wood, metal) frame with a drywall / siding construction (e.g., exterior walls) or a drywall / drywall construction (e.g., interior walls). However, this configuration is not required, and the walls 132 may have a different configuration (e.g., brick, fabric, glass, plastic, metal, lattice, trellis, cage, or log). It should be noted that the doors 134 or windows 136 may be omitted.

[0032] As shown in FIG. 1 , area 100 includes bathroom 118 defined by wall 128, wall 132, and door 138 and is shaped as a rectangle. However, it should be noted that bathroom 118 can be defined by wall 128, wall 132, or door 138 as any other suitable shape (e.g., square, oval, polygon, open-ended, closed-ended, teardrop, cornerless). Wall 128 includes a stud frame having a drywall / siding construction (exterior wall) or a drywall / drywall construction (interior wall). However, this configuration is not required, and wall 128 can be of a different configuration (e.g., brick, fabric, glass, plastic, metal, lattice, trellis, cage, or log). It should be noted that door 138 may be omitted.

[0033] Although area 100 includes bathroom 118, this is not required. For example, bathroom 118 could be omitted, or area 100 could be bathroom 118. Similarly, area 100 may be embodied as a residence for object 104, and the residence may have bathroom 118, a living area, and a kitchen area, but this is not required. As such, area 100 may be embodied as any residential area suitable for object 104 to live in. For example, area 100 may be embodied as a living room or living area, a dining room or dining area, a bedroom or sleeping area, a bathroom or bathroom area, a shower room or shower area, a playroom or play area, a home office or home office area, a basement or basement area, a garage or garage area, a storage room or storage area, an attic or attic area, an exercise room or exercise area, a dirt floor or dirt floor area, a closet or closet area, or any other suitable residential room or area, although non-residential areas may also be used. Similarly, although area 100 is shown as a residential area of object 104, this is not required. As such, area 100 may be embodied in a variety of ways. For example, area 100 may be embodied within or be a building, an apartment building, a single-family home, an accessory dwelling, a warehouse, a lobby, an office space, a cubicle, a hallway, a foyer, a hotel, a tent, a shed, a cage, a medical facility, a nursing home, a hospice, an assisted living facility, a hospital, a passenger area of a vehicle, a driver area of a vehicle, a control area of a vehicle, an elevator, a cockpit of an airplane or helicopter, a passenger cabin of an airplane or helicopter, a room of a boat, a cockpit or cabin of a boat, or any other suitable area.

[0034] Area 100 and bathroom 118 have various everyday life objects disposed therein, whether fixed (e.g., electrical appliances, plumbing fixtures) or movable (e.g., floor lamps, vases). These objects include lamp 106, sofa 108, table 110, chair 112, oven 114, refrigerator 116, bathroom 118, toilet 120, bathtub 122, and sink 124, any or all of which may be omitted from area 100. It should be noted that how these everyday life objects are arranged within area 100 is exemplary, and other layouts of these everyday life objects are possible.

[0035] Area 100 includes a floor, a ceiling, and corners, although the ceiling or corners may be omitted. Near the corners are stands 106 (e.g., tables, coffee tables, nightstands, chairs, shelves) on which devices 102 rest, are placed, or are positioned (e.g., stationary, fixed). However, it should be noted that stand 106 may be omitted or not, and device 102 may rest on or be attached to (e.g., fastened, engaged, or secured) the floor, or may be attached to (e.g., fastened, engaged, or secured) the ceiling or may be suspended from (e.g., via a cable or chain), or may be attached to (e.g., fastened, engaged, or secured) the wall 132 or wall 128 or may be suspended from (e.g., fastened, engaged, or secured) the wall 132 or wall 128.

[0036] The device 102 includes a processor (e.g., a controller, an edge processor, a single-core processor, a multi-core processor, a system-on-chip, a graphics processing unit, a hardware accelerator, a neural network accelerator, a machine learning accelerator) and a radar (e.g., a time-of-flight radar, a Doppler radar), where the processor is coupled (e.g., electrically, logically, mechanically) to the radar to control the radar (e.g., receive tracking data). For example, the processor may include a controller and a hardware accelerator. For example, the processor may enable local computing or edge computing to improve processing speed or provide data privacy or data security. The radar may have the set of components and field of view 130 shown in FIG. 7. For example, the field of view may be or include approximately 120 degrees (or less, or two or more angles) horizontally and approximately 90 degrees (or less, or more angles) vertically. Similarly, for example, the radar may be as disclosed in U.S. Pat. No. 9,019,150.

[0037] Although device 102 is shown in FIGS. 1-9 as having a housing (e.g., a container, enclosure, box, cube, cuboid, truncated pyramid, truncated cone, sphere, oval, television set, sound bar, speaker, bookend, flower pot, planter pot, vase, furniture, table, chair, sofa, bed, crib, shelf, bookshelf, television stand, appliance, dishwasher, refrigerator, overcoat, stove, toy, exercise equipment, treadmill, rowing machine, musical instrument, equipment) that hosts the processor and radar (e.g., internally, externally), this is optional and the housing can be omitted or modified. For example, several form factors of the housing are shown in FIG. 9. Similarly, for example, one or both of the processor and the radar may be housed in no housing at all or in different housings (e.g., the processor is housed in a first housing and the radar is housed in a second housing), and these different housings may be attached to each other, separated from each other, spaced apart from each other, facing each other, or in any other suitable configuration, and whether the housings are structurally or functionally identical or non-identical to each other.

[0038] As shown in FIG. 1 , device 102 includes a power line (e.g., a wire, cord, cable) through which power is supplied to the processor and radar. If device 102 includes other components as disclosed herein, these components may also be powered via the power line. As such, the power line includes an electrical plug, and device 102 is positioned in area 100 near outlet 126 such that the electrical plug is sufficiently slender or flexible to be plugged into outlet 126 and thereby power device 102. Outlet 126 may be 110 volts, 220 volts, or other voltage suitable for operating device 102. However, as disclosed herein, if device 102 includes a battery, which may be rechargeable, that is sufficiently energized to power the processor, radar, and any other components of device 102 as needed for a preset period of time (e.g., 30 minutes, 90 minutes, 120 minutes, 24 hours, 72 hours), the power line may be omitted or may be another power option.

[0039] The processor can activate the radar to operate within the area 100 and track the living object 104 within the area 100 when the object 104 is located within the field of view 130 within the area 100. The radar generates a set of data based on tracking the living object 104 within the area 100 when the object 104 is located within the field of view 130 within the area 100, and the processor determines based on the set of data whether the object 104 is experiencing an event (e.g., a medical emergency, fall, death, heart attack, stroke within the area 100 based on the set of data) and takes action (e.g., initiates communication with a remote telephone unit or server) corresponding to the event determined to be occurring within the area 100. For example, the processor can distinguish between a fast fall and a slow fall, each associated with a unique signature for medical purposes. For example, the event can be a medical event, a diagnostic prediction, or a diagnostic prognosis. For example, an action can be triggered by a threshold based on one or more criteria related to the object 104 being tracked by the radar or its environment. For example, the action can be (a) the device 102 (e.g., via its SIM module) calling a pre-configured phone number (e.g., a family member, caretaker, social worker, medical professional, nurse, personal doctor, medical facility, emergency services), (b) sending a message (e.g., via a Wi-Fi interface) to a server remote from the area 100, the device 102, and the object 104, (c) executing a set of pre-programmed escalation actions (e.g., sounding an "Are you OK?" message and calling a pre-configured phone number if there is no response from the object 104), or (d) other suitable action. As mentioned above, the object 104 does not have to reside in the area 100 to be within the field of view 130. In this manner, the object 104 can be present in the area 100 (e.g., whether in a repeat pattern, a one-time visit, for a relatively long period of time, or temporarily) and within the field of view 130.

[0040] The radar can operate in the Ku band, which includes values between about 12 GHz and about 18 GHz, the K band, which includes values between about 18 GHz and about 27 GHz, or the Ka band, which includes values between about 26.5 GHz and about 40 GHz, while complying with local radiation laws (e.g., regulated by the Federal Communications Commission) and without radiating radiation to other objects (e.g., stationary or mobile medical devices, wearable medical devices, pacemakers, insulin pumps, infusion pumps, microwave ovens, televisions, radios, Wi-Fi, cell phones, printers, and network equipment) in the operating vicinity. If the radar operates in at least two of these bands, the radar can be a single radar unit operating in at least two of these bands, or the radar can have at least two radar units, each dedicated to at least two of these bands. For example, the radar can operate in the Ku band, which includes values between about 12 GHz and about 18 GHz. For example, the radar can operate in the K band, which includes values between about 18 GHz and about 27 GHz. For example, the radar may operate in the Ka band between and including about 26.5 GHz and about 40 GHz.For example, the radar may operate in at least two of the Ku band, which is between about 12 GHz and about 18 GHz inclusive, the K band, which is between about 18 GHz and about 27 GHz inclusive, or the Ka band, which is between about 26.5 GHz and about 40 GHz inclusive, and may be continuously (e.g., to enhance resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation), such as when required by the processor based on the processor determining whether various criteria, signatures, or thresholds are met or not met, as disclosed herein. the radar may be switched between at least two of these bands to enhance resolution of the area 100 or its contents, or to manage power or heat dissipation), or parallel (e.g., operating on two of these bands simultaneously without self-interference for complementary or validating or confirming purposes), such as when required by the processor based on the processor determining whether various criteria, signatures, or thresholds are met or not met (e.g., to enhance resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation), as disclosed herein.For example, the radar may operate in the Ku band, between about 12 GHz and about 18 GHz, inclusive; the K band, between about 18 GHz and about 27 GHz, inclusive; or the Ka band, between about 26.5 GHz and about 40 GHz, inclusive, and may be configured to operate continuously (e.g., to detect the presence of an object 104 or area 100 or its contents), based on the processor determining whether various criteria, signatures, or thresholds are met or not, as disclosed herein (e.g., to enhance resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation). The radar may be parallel (e.g., operating simultaneously on two of these bands without self-interference for complementary or validating purposes), such as when requested by a processor based on the processor determining whether various criteria, signatures, or thresholds are met or not (e.g., to increase the resolution of the object 104 or the area 100 or its contents, or to manage power or heat dissipation), or parallel (e.g., operating on two of these bands simultaneously without self-interference for complementary or validating purposes), such as when requested by a processor based on the processor determining whether various criteria, signatures, or thresholds are met or not (e.g., to increase the resolution of the object 104 or the area 100 or its contents, or to manage power or heat dissipation), as disclosed herein. For example, the radar may switch frequencies within the Ku band between about 12 GHz and about 18 GHz, inclusive, such as when requested by a processor based on the processor determining whether various criteria, signatures, or thresholds are met or not (e.g., to increase the resolution of the object 104 or the area 100 or its contents, or to manage power or heat dissipation), as disclosed herein. For example, the radar may switch frequencies within the K-band between about 18 GHz and about 27 GHz, inclusive, as disclosed herein, such as when requested by the processor based on the processor determining whether various criteria, signatures, or thresholds have been met or not (e.g., to enhance resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation).For example, the radar may switch frequencies within the Ka band between and including about 26.5 GHz and about 40 GHz, such as when required by a processor based on the processor determining whether various criteria, signatures, or thresholds have been met or not (e.g., to improve resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation), as disclosed herein. These bands have unexpectedly proven to be technologically beneficial for a variety of reasons, as disclosed herein.

[0041] With respect to Ku-band, the radar may operate within the area 100 at radio frequencies between about 12 GHz and about 18 GHz inclusive (wavelengths between about 24.00 millimeters and about 16.65 millimeters), or the radar may switch frequencies within the Ku-band between about 12 GHz and about 18 GHz inclusive, for example, as required by a processor based on the processor determining whether various criteria, signatures, or thresholds have been met or not (e.g., to enhance resolution of objects 104 or the area 100 or its contents, or to manage power or heat dissipation, as disclosed herein). It has been unexpectedly discovered to be beneficial in the art that this band penetrates walls / objects within the field of view 130 better than higher frequencies and allows the radar to reach further, useful for tracking the position of objects 104 within the area 100 and detecting the attitude of objects 104 within the area 100.

[0042] With respect to the K-band, the radar operates within the area 100 in a radio frequency band between and including approximately 18 GHz and approximately 27 GHz (wavelengths between approximately 16.65 millimeters and approximately 11.10 millimeters), or the radar may switch frequencies within the K-band between and including these values (e.g., to enhance resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation, as disclosed herein), such as when requested by a processor based on the processor determining whether various criteria, signatures, or thresholds have been met or not. This band has unexpectedly been found to be technologically beneficial because it corresponds to the peak of the absorption spectrum of water. This is because, in some situations, conventional radars may not be configured to operate at frequencies above 22 GHz because those frequencies are readily absorbed by water. Therefore, conventional radars typically desire their signal to penetrate water vapor to reach other targets that are more reflective (e.g., metal). Thus, if the object 104 is a mammal such as a human, which may have a water content of up to 60%, a radar tracking the object 104 within the field of view 130 can produce a large, measurable change in the amount of reflected signal within the area 100, which increases the accuracy or precision of the radar operating within the area 100 and tracking the object 104 within the field of view 130 within the area 100. Within the K-band, as explained above, the radio frequency ranges between about 23 GHz and about 25 GHz, and particularly between about 24 GHz and including these values, have been unexpectedly beneficial.

[0043] With respect to the Ka band, the radar operates within the area 100 within a radio frequency band (wavelengths between approximately 11.31 millimeters and approximately 7.49 millimeters) between and including approximately 26.5 GHz and approximately 40 GHz, or the radar may switch frequencies within the Ka band between and including these values (e.g., to enhance resolution of the object 104 or the area 100 or its contents, or to manage power or heat dissipation, as disclosed herein), such as when requested by a processor based on the processor determining whether various criteria, signatures, or thresholds have been met or not. This band has unexpectedly been found to be technically beneficial for facilitating detection of vital signs (e.g., heart rate, breathing rate) located within the field of view 130 when the object 104 is a mammal, such as a human. Because the object 104 may repeatedly deflect within the field of view 130 due to inhalation / exhalation of oxygen or pulsation of blood, this condition results in a greater percentage change in the wavelength of the carrier frequency detected by the processor.

[0044] The radar may switch modalities between Doppler mode (or another radar modality) and time-of-flight mode (or another radar modality) when requested by the processor based on the processor's determination of whether various criteria, signatures, or thresholds are met or not met (e.g., to enhance resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation), as disclosed herein. Note that such switching may or may not operate in series or parallel, may or may not interfere with each other, and may or may not be accompanied by frequency switching or band switching, whether the radar is operating in Ku band, K band, Ka band, or other band, as disclosed herein. For example, a radar may have a first radar unit operating in Doppler mode and a second radar unit operating in time-of-flight mode, with the processor requesting that the first radar unit operate in Doppler mode and then switch to the second radar unit to operate in time-of-flight mode, or vice versa, based on the processor determining whether various criteria, signatures, or thresholds have been met or not (e.g., to enhance resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation, as disclosed herein, although parallel or serial radar mode operation is also possible). Note that the first and second radar units may be hosted by a common housing (e.g., internally, externally), or each may have its own housing that may be spaced apart (e.g., within about 5, 4, 3, 2, or 1 feet or meters) from one another, as disclosed herein.Similarly, for example, the radar may be capable of parallel or serial radar mode operation, but may operate in Doppler or time-of-flight mode, where the processor requests the radar to operate in Doppler mode and then switch to time-of-flight mode, or vice versa, based on the processor determining whether various criteria, signatures, or thresholds have been met or not (e.g., to enhance resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation), as disclosed herein.

[0045] This radar is designed for approximately 5 GHz (and each band), approximately 24 GHz (and each band), approximately 60 GHz (and each band), and other operating frequencies (all and each band work, but some work better than others for some use cases), with experience from many prototypes including radar, signal processing, and artificial intelligence expertise. For example, for approximately 5 GHz (and each band) or approximately 60 GHz (and each band), the radar can operate within these frequencies or their corresponding bands, switch frequencies within them, or switch bands with Ku-band, K-band, Ka-band, or other bands disclosed herein. In some embodiments, some design parameters relate to the field of view (the left and right limits of what the radar can see). For example, the field of view can be approximately 120 degrees (or below or above) horizontally, approximately 90 degrees (or below or above) vertically, or other. In some embodiments, design parameters relate to the resolution (the granularity with which the radar can distinguish details within the field of view). Resolution can be achieved through voxels (3D pixels) with a "width" of approximately 15 degrees, a "height" of approximately 15 degrees, and a depth of approximately 25 cm, useful for location and fall detection, although individually higher or lower resolution forms are possible. It is useful for measuring heart rate, respiratory rate, or other vital signs. In some embodiments, some design parameters relate to penetration (the balance between the radar's ability to penetrate general objects (walls, furniture, etc.) and reflection from monitored objects (such as humans)). For example, some embodiments enable penetration of US-standard studded drywall or two layers of drywall / siding at approximately 20 meters with good or sufficient reflection from human targets. Note that this distance is exemplary and can be increased or decreased based on other parameters (e.g., supplemental data sources, supplemental radar, type of material used to manufacture walls or appliances or furniture, height of radar relative to ground floor or physical area or monitored floor or physical area, power limitations set by government authorities).In some embodiments, the design parameters relate to transmit power (where the field of view defines at least some of the directions the radar scans, transmit power affects how far the radar can see).

[0046] As explained above, radars operating in the K-band, which includes wavelengths between about 18 GHz and about 27 GHz (wavelengths between about 16.65 mm and about 11.10 mm), particularly between about 23 GHz and about 25 GHz, and more particularly between about 24 GHz and about 18 GHz, have been unexpectedly beneficial due to their good balance between signal penetration, tracking distance, and human detection, while being subject to regulatory power limits set to stay well within human safety parameters across a very wide range of applications or use cases, as disclosed herein. Operating radars in the K-band, which includes wavelengths between about 18 GHz and about 27 GHz (wavelengths between about 16.65 mm and about 11.10 mm), particularly between about 23 GHz and about 25 GHz and about 24 GHz, more particularly at about 24 GHz, allows the radar to further adjust the field of view and resolution by modifying certain characteristics of some antenna arrays or by adding additional radar subsystem boards that may be required for future applications. For example, this mode of operation allows for smaller antennas and a more compact overall antenna array, while still allowing for relatively high frequencies that enable accurate distance measurements, and as disclosed herein, allows for integration into a variety of form factors.

[0047] Although the material penetration characteristics of radar operating in the K band, between about 18 GHz and about 27 GHz inclusive (wavelengths between about 16.65 millimeters and about 11.10 millimeters), particularly between about 23 GHz and about 25 GHz inclusive, and more particularly at about 24 GHz, are much better for tracking indoors (e.g., within area 100) than operating at about 60 GHz or about 76-78 GHz, as disclosed herein, operation at about 60 GHz (or a corresponding band, or a switch frequency therein, or a switch band with Ku band or K band or Ka band or other bands disclosed herein) or about 76-78 GHz (or a corresponding band, or a switch frequency therein, or a switch band with Ku band or K band or Ka band or other bands disclosed herein) may be sufficient for some use cases indoors (e.g., within area 100). Specifically, radars operating within the K-band, which covers frequencies between approximately 18 GHz and approximately 27 GHz (wavelengths between approximately 16.65 mm and approximately 11.10 mm), particularly between approximately 23 GHz and approximately 25 GHz, and more particularly at approximately 24 GHz, can operate through several layers of standard wall construction and see through various types of clothing. For example, some radars above 60 GHz are limited in their ability to operate indoors because walls may be substantially opaque or may be strongly affected by noise-generating clothing. For example, radars operating at approximately 60 GHz can detect the heart and breathing rates of a relatively stationary object 104 standing within approximately 7 meters in front of the radar. However, some radar embodiments operating at approximately 60 GHz do not penetrate solid objects, including clothing and the human body, well enough. Therefore, measurements may be noisy if the object 104 moves and flap clothing, if any. In contrast, radar operating within the K-band, between about 18 GHz and about 27 GHz (wavelengths between about 16.65 millimeters and about 11.10 millimeters), inclusive, particularly between about 23 GHz and about 25 GHz, inclusive, and more particularly at about 24 GHz, can penetrate human bodies, thereby minimizing motion-related noise. Furthermore, when radar operates based on voxel-based tracking within the K-band, between about 18 GHz and about 27 GHz (wavelengths between about 16.65 millimeters and about 11.10 millimeters), inclusive, particularly between about 23 GHz and about 25 GHz, inclusive, and more particularly at about 24 GHz, the processor can discard, remove, delete, or ignore any or certain voxels that do not intersect with an object 104, such as a human body, or alternatively, can simultaneously measure the vitals of multiple objects 104, such as human bodies, within the area 100.For example, when using voxel-based tracking, a processor receives a set of data from the radar, accesses a set of voxels formed based on the set of data, discards, removes, deletes, or ignores a first subset of voxels from the set of voxels based on the first subset of voxels not representing living objects 104 in the area 100, so that a second subset of voxels is identified from the set of voxels, and takes an action as disclosed herein in response to an event determined to be occurring within the defined area based on the second subset of voxels. Therefore, this form of voxel filtering allows for greater processing efficiency. Furthermore, radar operating within the K-band, between about 18 GHz and about 27 GHz (wavelengths between about 16.65 millimeters and about 11.10 millimeters), including these values, particularly between about 23 GHz and about 25 GHz, including these values, and more particularly at about 24 GHz, allows for millimeter-level resolution within voxels, allowing for the visualization of an individual's tissue and skin movement as a function of heart rate and respiration.

[0048] When the device 102 is provided to a user, who may or may not be an object 104 (e.g., DIY) (e.g., a proxy, child, or caretaker acting on behalf of a parent or care recipient), the user may receive instructions regarding (e.g., configuration and use) the device 102. For example, the user may be an object (e.g., DIY) or may not be an object (e.g., a proxy, child, parent, or caretaker acting on behalf of a care recipient). The device 102 may be provided to the user in a variety of ways. For example, the device 102 may be mailed (e.g., US Mail), couriered (e.g., FedEx), shipped (e.g., in a package), sent, handed over, delivered, present or installed in a community, residence, or vehicle, or otherwise suitably available to the user. The user may be instructed in a variety of ways. For example, the user may be instructed via a medium such as a written, diagrammed, or printed manual, a computer file (e.g., a portable document format file, a presentation file, a word processing file, an image file, a video file, an audio file), a website, a mobile application, an audio or pictorial guide, an auditory or visual wizard, a call center, or any other suitable instructions regarding the device (e.g., how to use it). The user may be instructed as to how and where to position the device 102 within the area 100 in which the object 104 may reside (e.g., relatively temporarily). The user may be instructed as to how to power, turn on, and activate the radar within the area 100. If initial configuration or setup is required, the user may be instructed accordingly.

[0049] Once the radar is configured and activated, it tracks objects 104 living (or located) within the area 100, generates sets of data based on tracking the objects 104 living (or located) within the area 100, transmits the sets of data to a processor, and the processor determines whether to take action based on the set of data (e.g., individually or in combination with other knowledge, predictions, estimates, inferences, or data from data sources regarding the objects 104 or the area 100), and takes action based on the set of data. For example, the processor can determine based on the set of data whether the objects 104 are experiencing an event (e.g., a medical emergency, fall, death, heart attack, stroke, diagnostic prediction, diagnostic inference, diagnostic prediction) within the area 100, and take action (e.g., initiate communication with a remote telephone unit or a remote server) in response to the event determined to be occurring within the area 100. For example, the processor may form a signature of the object based on a set of possible data over a period of time, compare the signature to a set of signature templates corresponding to a set of events (e.g., medical emergency, fall, death, heart attack, stroke, diagnostic prediction, diagnostic prediction), and then determine whether the object 100 is experiencing an event based on determining whether a match threshold between the signature and the set of signature templates is met or not. For example, the match threshold may be met or not to infer that the health status (e.g., activities of daily living, amount of exercise, speed of movement, reaction time) of the object 104 may be deteriorating.

[0050] An event can be related to an activity of daily living (e.g., eating, drinking, sleeping, washing, bathing, toileting, reading, sitting, exercising, laundry, cooking, cleaning) of an object 104 within the area 100. For example, the event can be identifying that the object 104 is performing an activity of daily living adequately or inadequately, or that there is a decrease, increase, or maintenance in the number, frequency, or quality of the activity of daily living. Similarly, the event can be related to the object 104 falling within the area 100. Similarly, the event can be related to the object 104 being motionless within the area 100 for a predetermined period of time (e.g., dead, dying, paralyzed, injured, unconscious, sleeping). Furthermore, the event can be related to the object 104 being absent from the area 100 for a predetermined period of time (e.g., lost, missing, injured, dead, unconscious, seizure, obstruction). And this event can relate to the object 104 not being tracked within the area 100 for a predetermined period of time while the object 104 is within the area 100 (e.g., dead, paralyzed, injured, unconscious, sleeping in an occluded area or coverage gap).

[0051] Note that the processor need not determine whether objects 104 in area 100 are experiencing an event. Thus, whether additionally or alternatively, the processor determines whether an action should be taken based on a set of data, which may be event-independent, event-dependent, or event-free, and takes an action based on a set of data, which may be event-independent or event-independent. For example, the processor may determine, as disclosed herein, that there is no event being experienced by the object 104 in the area 100, or that a frequency change within a band (e.g., Ku band, K band, Ka band) may be needed (e.g., to increase the resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation) independent of or agnostic to the event, or that a band change (e.g., between at least two of the Ku-band, K-band, or Ka-band) may be needed (e.g., to increase the resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation) independent of or independent of the event, or that a set of data may need to be validated or confirmed by another data source (e.g., a set of microphones, cameras, sensors) independent of or independent of the event.

[0052] As described herein, devices 102 may be embodied in or as a wide variety of everyday household objects, such as televisions, dumb and smart speakers, bookends, flower pots, planter pots, vases (or other containers), furniture (e.g., tables, chairs, sofas, wall-mounted or tabletop or shelf-mounted picture frames, bookcases, shelves, dressers, beds, cabinets, shutters), and appliances (dishwashers, ovens / cooktops, microwaves, refrigerators, washers, dryers, etc.). Such implementations provide distinct technological enhancements to the various sensing capabilities of current offerings, as disclosed herein. However, it should be noted that additional or alternative non-domestic or non-residential use cases are also possible. These may include applications in cars and vehicles (such as buses, boats, planes, and railcars) for measuring eye movement, facial expression, mouth movement, head movement, joint movement, heart rate, respiratory rate, or other vitals while people are driving, commuting, or traveling to other destinations, and as people live, work, or play. For example, if the device 102 is installed in or included in a vehicle (e.g., car, truck, bus, cockpit), the radar as disclosed herein can track whether the vehicle operator (e.g., vehicle driver) or passenger is tired, drowsy, asleep, unconscious, dead, having a seizure, or intoxicated, and then take appropriate programmed action (e.g., not starting the vehicle, slowing or stopping the vehicle if safe, notifying an emergency number or radio contact from the vehicle, notifying a preset number or radio contact from the vehicle, sending a geolocation from the vehicle to a remote server, opening a vehicle window, sounding the vehicle horn or security alarm, opening a door if it is safe or too cold or the vehicle is not moving, posting to social media from the vehicle).Similarly, when device 102 is installed in or contained within a vehicle (e.g., a passenger car, automobile, van, minivan, sport utility vehicle), the radar, as disclosed herein, can determine whether an infant (or infant, toddler, disabled, infirm, or elderly person) is in a child safety seat in the vehicle (e.g., if no other person is being tracked in the vehicle other than the infant (or infant, toddler, disabled, infirm, or elderly person)) and take appropriate programmed actions (e.g., starting the vehicle, turning on the vehicle's air conditioning or heating or air conditioning, notifying an emergency number from the vehicle, notifying a pre-set number from the vehicle, transmitting location information from the vehicle to a remote server, opening a vehicle window, sounding the vehicle's horn or vehicle security alarm, opening a vehicle door if it is safe or not too cold and the vehicle is not moving, posting to social media from the vehicle). Tracking of these situations can be supplemented by data from other input devices (e.g., microphones, cameras, sensors) and powered by an on-board battery, as disclosed herein.

[0053] As described herein, a radar operating in the K-band, between about 18 GHz and about 27 GHz, inclusive (wavelengths between about 16.65 millimeters and about 11.10 millimeters), particularly between about 23 GHz and about 25 GHz, inclusive, and more particularly about 24 GHz, can be a time-of-flight radar that provides a set of data to a processor for the processor to convert into a set of voxels to simulate an object 104, which can include simulating an area 100. The radar, as disclosed herein, can switch bands (e.g., between Ku-band, K-band, or Ka-band) or frequencies (e.g., within Ku-band, K-band, or Ka-band) to increase the resolution of the set of voxels, whether incomplete, imprecise, inaccurate, or meeting or not meeting a preset resolution threshold. In this manner, the radar can track the location, posture, or vital signs of objects 104 (e.g., humans, elderly, disabled, or infirm) within the field of view 130, whether above or below ground, indoors or outdoors, in homes; hospitals or nursing homes or rehabilitation facilities; senior living facilities; care homes and shelters, or other suitable medical or non-medical facilities.

[0054] As described further herein, a set of microphones included in device 102 and controlled by a processor can optionally provide high-quality 3D audio mapping to locate or localize sound sources and can also optionally provide two-way audio capabilities. For example, the set of microphones can feed audio data (and other sensors with their respective data) to a processor to enable a flexible computing platform that combines hardware-accelerated artificial intelligence (AI) and processing at the edge, i.e., edge computing. As described further herein, device 102 can include single or various communication options (e.g., wired, wireless, waveguide) so that device 102 can act as a hub for other devices and transfer data, alerts, and important information to and from virtual computing instances (e.g., servers, applications, databases) running within cloud computing services (e.g., Amazon Web Services, Google Cloud) or phones, whether wired, wireless, or waveguide-based. To maximize efficiency and security, device 102 can be securely and periodically updated over the air to support various processor / radar algorithm improvements and additions, as well as an optionally evolving ecosystem of devices and services.

[0055] As described herein, the radar transmits a set of data (e.g., raw data) to a processor (e.g., an edge processor), which processes the data with on-board algorithms to reconstruct high-resolution 3D information (e.g., voxel information) about the area 100 within the field of view 130. The device 102 then applies various techniques to determine or infer various information about the object 104. For example, if the object 104 is a human, the processor may determine the body position of the object 104 (e.g., standing, lying down, supine, prone, supine) or the object's 104 activities of daily living (e.g., sitting, standing, walking, falling, sleeping, eating, exercising, showering, bathing, cooking). The processor combines this information with data from other sensing subsystems (e.g., audio, vision, temperature, motion, distance, proximity, moisture) and additional data sources, if available, ranging from fitness apps (e.g., tracking the object 104) to social media (e.g., object 104 profile) to electronic health records (e.g., object 104 profile) to know (e.g., deterministically) rather than guess (e.g., probabilistically) what condition people are in, allowing the electronic device 102 to enable people's safety, well-being, and supported self-care while minimizing false alarms and missed signs of increased risk.

[0056] As described herein, device 102 benefits from a virtuous cycle of growing data sets. These data sets drive improved quality and increased user value, resulting in more usage and more data. In some embodiments, device 102 comprises a multi-sensor ambient sensing platform, and cloud computing services provide valuable context. For example, device 102 is further developed / enhanced through ongoing in-field A / B testing (e.g., over-the-air firmware updates) to generate improvements in sensor processing, health monitoring, and user experience, all based on continuously improving AI, catalyzing additional features, use cases, and resulting offerings over time. For example, device 102 may include processor-accessible memory (e.g., persistent memory, flash memory) that stores AI models (e.g., based on convolutional neural networks, recurrent neural networks, reinforcement algorithms) trained to determine whether objects 104 meet or fail various criteria, signatures, or thresholds that may relate to the health, safety, or security of objects 104 within area 100. In this manner, if the subject 104 is a human, the processor can access the AI model to analyze a baseline over time and observe early indicators of changes in the human's safety, well-being, and health. Thus, using the device 104 to track the subject 104 over time should also improve monitoring based on the AI model's better understanding of the subject 104.

[0057] As described herein, there are many exemplary use cases for device 102. For example, if subject 104 is a human, observing (and having a processor act upon) daily biometric measurements, movements, and activities of daily living (e.g., by a processor via radar) can help proactively detect potential emerging health and safety risks well before episodic physical, laboratory, and diagnostic tests, and sometimes well before awareness of deteriorating health. Through observing (e.g., by a processor via radar) and analyzing (e.g., by a processor via radar) daily behavioral patterns, such as micro-actions like eating, drinking, sleeping, exercising, washing, bathing, showering, toileting, socializing, getting dressed, and taking medication, a digital picture of the person's health and well-being emerges (on which the processor can act). Furthermore, by observing (e.g., by a processor via radar), analyzing, and acting on (e.g., electronically notifying the individual, family, or physician via sound prompts, warning messages, or a mobile app) movement, time in bed, and frequency of toilet use, the device 102 can indicate signs of illness that may not be readily apparent to the individual, family, or even a clinician. Thus, not only can the device 102 increase awareness of such undetected disease states, but the device 102 can also help identify (e.g., by a processor) deterioration in mental or physical function and monitor (and act on) signs of worsening chronic conditions already experienced by the person. The device 102 enables detection (e.g., by a processor) of such signs of risk to people's safety, well-being, and health long before they result in more acute symptoms and ambulance dispatches, emergency room admissions, and hospitalizations. The desired end result, therefore, is reduced suffering for the people supported by the device 102, reduced costs through proactive intervention, and efficient use of precious medical resources.

[0058] One use case of the device 102 is to detect a fall of an object 104 via the processor. For example, if the object 104 is an elderly human, the fall may be a health and safety hazard. While falls are common, this general term hides significant variability and subtlety in the events suffered by the object 104. The location, circumstances, direction, speed, bearing, angle, or other characteristics and subtleties of the fall tracked by the radar and transmitted to the processor may provide clues to the processor as to its underlying cause, and these details have significant clinical implications. Unfortunately, many people who fall sustain injuries that complicate efforts to determine what actually happened, often presenting physicians with diagnostic challenges. The speed or direction of the fall, as determined by the processor based on a set of data from the radar, may affect the likelihood of head injury, fracture, and recovery. In this way, the processor can consider or detect various fall factors based on a set of data from the radar. These factors include location (which room, location within the room, type, and strength). Such factors include situations such as showering versus entering a room with a threshold. These factors can include outcome - stuck, tripping, or movement without standing. These factors can include a fall out of bed, a fall from a standing position, or a tilt from a chair. These factors can include directionality - a simple trip or possibly correlated with an event affecting the brain. These factors can include speed (speed increases injury) and trajectory indicates a fall or tipping. Such factors can include disability, and asymmetric movement may indicate fractures. Thus, some of these factors may be related to the severity of the injury or the presence or absence of an underlying condition.

[0059] One use case of device 102 is to detect functional limitations of object 104 via the processor. For example, if object 104 is an elderly human, the functional limitations may be dangerous to health and safety. A frail human may experience various changes in mental and physical function over time (e.g., as tracked by radar and detected by a processor) before becoming impaired enough to require care or be noticed by others, whether through cognitive or physical limitations. Detectable evidence of a decline in mental function (e.g., as tracked by radar and detected by a processor) may include changes in self-care patterns (e.g., as tracked by radar and detected by a processor) or the ability to maintain sleep (e.g., as tracked by radar and detected by a processor). Evidence of decreased physical function (e.g., as tracked by radar and detected by a processor) may include a decrease in overall physical activity (e.g., as tracked by radar and detected by a processor), a decrease in walking speed (e.g., as tracked by radar and detected by a processor), or evidence of stumbling (e.g., as tracked by radar and detected by a processor).

[0060] One use case of device 102 is to detect changes in activities of daily living via a processor. For example, if object 104 is an elderly human, such changes may be hazardous to health and safety. Elderly individuals are routinely assessed for their ability to perform activities of daily living (e.g., tracked by radar and detected by a processor), and these changes (e.g., tracked by radar and detected by a processor) are part of the aging process that may correlate with future problems in self-care, coping, and independence. Along with disease states (e.g., tracked by radar and detected by a processor), these changes may be an important part of a Frailty Index (e.g., tracked by a processor), which is a proxy for vulnerability to poor outcomes. Thus, device 102 can detect human engagement (e.g., as tracked by radar and detected by a processor) in several major activities of daily living, including sleeping, eating, drinking, toileting, socializing, or others. For example, based on a set of data received from the radar, the processor can track sleep and detect sleep interruptions, schedule changes, time to restful sleep, or early awakening. For example, based on a set of data received from the radar, the processor can track the timing, frequency, preparation time, eating and drinking time, or whether cooking is performed. For example, based on a set of data received from the radar, the processor can track the timing and frequency of toilet use. For example, based on a set of data received from the radar, the processor can track sociability - measuring time spent alone, time spent on the phone, frequency and duration of visits by others, and number of visitors present. As described herein, the processor can distinguish objects 104 from others. For example, based on a set of data received from the radar, the processor can track microbehaviors (e.g., medication use, dressing, grooming).For example, based on a set of data from a radar, which may be enhanced or supplemented (e.g., augmented) with data from other data sources disclosed herein, the processor may track the object 104 and another object 104 (e.g., ) and distinguish, based on the set of data, an object 104 located or living within the area 100 from another object 104 located or residing within the area 100, based on the set of data, before determining whether the object 104 is experiencing an event involving the object 104 within the area 100 (e.g., by learning the habits and signatures of the object 104 over time). For example, over time, once the processor is able to distinguish between the object 104 and another object 104 within the area 100, the processor may determine whether an event is custom (or unique) to the object 104 based on distinguishing between the object 104 and another object 104 based on the set of data.

[0061] One use case of device 102 is to detect evidence of illness via the processor. For example, if the subject 104 is an elderly human, there may be an illness dangerous to their health and safety. Many illnesses are evidenced by an abnormality that the processor can identify based on a set of data from the radar and call attention to (e.g., electronically notifying the subject, family, or physician via a sound prompt, warning message, or mobile app). The processor (e.g., running various AI algorithms) and cloud analytics logic (e.g., running within a cloud computing instance) enable correlation of various factors (sent by device 102 via the communications unit) that aid in recommending or suggesting at least some medical evaluation via an output device (e.g., speaker, display) of device 102 requested by the processor. Examples of these conditions include cardiopulmonary conditions - irregular heart rate (via heart rate analysis detected by the processor from the set of data transmitted by the radar), loss of cardiac reserve (increased heart rate above baseline during normal activity detected by the processor from the set of data transmitted by the radar), or loss of respiratory reserve (increased respiratory rate above baseline during normal activity detected by the processor from the set of data transmitted by the radar). Examples of these conditions include sleep disorders - nighttime awakenings (via atypical rises during the night detected by the processor from the set of data transmitted by the radar), snoring patterns (via microphone detected by the processor), sleep apnea (via analysis of respiratory rate and pauses detected by the processor from the set of data transmitted by the radar or microphone).Examples of these diseases include neurocognitive diseases - cognitive decline (abnormal repetition of behavior detected by the processor from the set of data transmitted by the radar), relapse patterns (detected by the processor from the set of data transmitted by the radar), skipping dishes (detected by the processor from the set of data transmitted by the radar), or decline in self-care (detected by the processor from the set of data transmitted by the radar). Examples of these diseases include neuromuscular diseases - movement disorders - via gait analysis (detected by the processor from the set of data transmitted by the radar), rise time analysis (detected by the processor from the set of data transmitted by the radar), and gait speed analysis (detected by the processor from the set of data transmitted by the radar), or seizure detection (abnormal movements detected by the processor from the set of data transmitted by the radar, abnormal movements preceding a fall, post-seizure state). Examples of these diseases include infectious diseases, such as the detection of a cough (detected by the processor from the set of data transmitted from the radar or microphone) may indicate the onset of influenza, pneumonia, upper respiratory infection, or COVID; the frequency of toileting or defecation (detected by the processor from the set of data transmitted from the radar) may indicate the presence of a urinary tract infection, diarrhea, or constipation. Examples of such diseases include metabolic diseases—frequency of toilet use (detected by the processor from the set of data transmitted by the radar) may indicate uncontrolled diabetes or urinary system problems or stomach problems, or thyroid conditions may subtly alert resting heart rate over time (detected by the processor from the set of data transmitted by the radar). For example, with respect to various diseases disclosed herein, the processor can be programmed to identify potential symptoms, onset, or signs of a particular disease (e.g., neurological disease, neurodegenerative disease, Parkinson's disease, Alzheimer's disease, facioscapulohumeral muscular dystrophy, Crohn's disease, chronic obstructive pulmonary disease, atopic dermatitis) based on the set of data from the radar.

[0062] One use case of the device 102 is to monitor (by a processor from a set of data transmitted by the radar) a chronic condition. For example, if the subject 104 is an elderly human, the chronic condition may be dangerous to health and safety. Various factors for successful management of a chronic disease are related to an individual's behavior. In addition to understanding whether the individual's activities are promoting health, the device 102 can detect (by a processor from a set of data transmitted by the radar) changes that suggest the individual is suffering from a worsening condition. There are various chronic conditions that can be monitored by the device 102 (processor). For example, some heart failures and rhythm disorders can be monitored by heart rate and respiratory rate (by a processor from a set of data transmitted by the radar), movement speed (by a processor from a set of data transmitted by the radar), physical activity time (by a processor from a set of data transmitted by the radar), or medication use (by a processor from a set of data transmitted by the radar). Some asthma and chronic obstructive pulmonary disease (COPD) may be monitored by heart rate and respiration rate (by a processor from a set of data transmitted by radar), cough and wheeze measurements (by a processor from a set of data transmitted by radar), movement speed (by a processor from a set of data transmitted by radar), or medication use (by a processor from a set of data transmitted by radar). Some sleep disorders may be monitored by sleep duration (by a processor from a set of data transmitted by radar), interruptions (by a processor from a set of data transmitted by radar), snoring detection (by a processor from a set of data transmitted by radar or a microphone), or heart rate and respiration rate (by a processor from a set of data transmitted by radar).Some neurobehavioral conditions can be monitored by tracking potential declines in daily activities (by a processor from the set of data transmitted by the radar), movement speed (by a processor from the set of data transmitted by the radar), medication use (by a processor from the set of data transmitted by the radar), or near-falls or falls (by a processor from the set of data transmitted by the radar). For example, a processor, based on the set of data from the radar or augmented by data from other sensors, as disclosed herein, can fuse such data to not only detect falls (or other events), but also estimate or predict at least some emotional health, mood, and stress level of the subject 104.

[0063] Figure 3 illustrates one embodiment of a device including a radar according to the present disclosure. Figure 4 illustrates an embodiment of a set of devices, each including a radar or sensor according to the present disclosure. In particular, device 102 is positioned on a stand 106. The device includes a housing having an upper side, a lower side, and a middle portion spanning between the upper and lower sides. Note that the top or middle portion may be omitted.

[0064] The top surface includes wood but can include other suitable materials (e.g., plastic, metal, rubber, fabric, glass). The bottom surface includes wood but can include other suitable materials (e.g., plastic, metal, rubber, fabric, glass). The middle portion includes fabric but can include other suitable materials (e.g., plastic, metal, wood, rubber, glass). Note that the top, bottom, and middle portions can be selectively interchanged to create different looks as desired, as shown in FIG. 4.

[0065] Although the housing has an oval shape when viewed from above, this is not required and the housing (or other form of support) may be shaped in other aspects (e.g., a container, enclosure, box, frame, base, cube, cuboid, pyramid, cone, sphere, oval, television apparatus, sound bar, speaker, bookend, flower pot, planter pot, vase, furniture, table, chair, sofa, bed, crib, shelf, bookshelf, TV stand, appliance, dishwasher, refrigerator, overcoat, stove, toy, exercise equipment, treadmill, rowing machine, musical instrument, fixtures, electrical equipment, plumbing equipment).

[0066] 5 is a logic diagram of one embodiment of a device including a radar according to the present disclosure. In particular, the device 102 includes a power supply unit (internal or external), a logic unit powered via the power supply unit, a housing (e.g., enclosure) that holds (optionally itself or any of its components) a communication unit (optionally itself or any of its components) that is controlled (any of its components) via the logic unit and powered (any of its components) via the power supply unit, a speaker unit (optionally itself or any of its components) that is controlled (any of its components) via the logic unit and powered (any of its components) via the power supply unit. a connector unit (optionally itself or any component thereof) controlled (any component) through the logic unit and powered (any component) through the power supply unit; a user interface unit (any itself or any component thereof) controlled (any component) through the logic unit and powered (any component) through the power supply unit; a microphone unit (any itself or any component thereof) controlled (any component) through the logic unit and powered (any component) through the power supply unit; and a sensor unit (any itself or any component thereof) controlled (any component) through the logic unit and powered (any component) through the power supply unit.

[0067] The logical units include a processor, random access memory (RAM), persistent storage (e.g., an embedded multimedia card, a solid-state drive, a flash memory), and a hardware accelerator (e.g., a neural network accelerator, a machine learning accelerator), although the hardware accelerator is optional. The communication units (either themselves or any components thereof) include a cellular SIM unit (e.g., a SIM card receiver and corresponding circuitry), a cellular modem, a Wi-Fi unit (e.g., a Wi-Fi chip), a Bluetooth (or another short-range wireless communication standard) unit (e.g., a Bluetooth chip), a Zigbee (or another short-range automation wireless communication standard) unit (e.g., a Zigbee chip), a Z-wave (or another short-range automation wireless communication standard) unit (e.g., a Z-wave chip), a near-field communication (NFC) unit (e.g., an NFC chip), and a registration jack (RJ). The communication units are used for various communications as disclosed herein. For example, a Bluetooth or Wi-Fi unit can be used when the device 102 needs to be paired (e.g., for configuration or setup) with a desktop or mobile device (e.g., phone, laptop, tablet, wearable computer) or another device 102 or sensor as disclosed herein. A Wi-Fi unit or RJ can be used when the device 102 needs to be connected to Wi-Fi (e.g., to send data generated by the processor to or receive data from a cloud computing service) or when it needs to be connected to a desktop or mobile device (e.g., phone, laptop, tablet, wearable computer) or another device 102 or sensor as disclosed herein.For example, if the device 102 has a lamp (e.g., hosted internally or externally by its housing), needs to turn on an indicator light (on its own or another device), needs to sound a general alarm (on its own or another device), or needs to monitor a property of the object 104 using another device (e.g., using another device, a Zigbee or Z-wave protocol (or other suitable automation communication protocol) can be used by a lamp (on its own or another device), an alarm (on its own or another device), a health monitoring device (on its own or another device), and general smart home automation (on its own or another device).) If the device 102 needs to be configured for various operations as disclosed herein, such configuration can be done via a mobile app running on a mobile phone, tablet computer, wearable computer, or another computing unit (e.g., laptop, desktop) operated by the object 104 or a user and communicating with the device 102, as disclosed herein. This arrangement simplifies the collection of configuration information (such as Wi-Fi passwords and authentication codes). Whatever computing unit is used can then use its NFC unit to transmit its configuration information (e.g., wirelessly) to device 102, allowing device 102 to use other communication standards. Note that Bluetooth communication technology (or other short-range communication protocols) may additionally or alternatively be used to transmit such configuration information (and the computing unit and device 102 may be configured accordingly). If device 102 needs to synchronize operation or exchange data with other devices 102 (e.g., two radars cooperatively scanning area 100), any communication component of the communication unit (e.g., Bluetooth unit, Z-wave unit, Zigbee unit, Wi-Fi unit, RJ unit) can be used. The speaker unit (itself or any component) may include an amplifier or speaker.The speaker unit is used for various sound, vocal, or audio outputs as disclosed herein. For example, configuration, health, safety, suggestion, recommendation, or wellness content may be output by the speaker unit as disclosed herein. The power supply unit may include a rear panel connector (optional), a barrel jack (optional), power distribution / power supply, a charge controller (optional), and a battery backup (optional). The user interface unit (optional itself or a component thereof) includes a physical interface (e.g., buttons, dials, levers, switches) and a virtual interface (e.g., a touch or non-touch display presenting a graphical user interface). The user interface unit is used for various tactile or visual outputs as disclosed herein. For example, configuration, health, safety, suggestion, recommendation, or wellness content may be output by the user interface unit as disclosed herein. The microphone unit (optional itself or a component thereof) may include a set of microphones. The microphone unit is used for various capture of sound, vocal, or audio input as disclosed herein. For example, as disclosed herein, a voice command or sound identified within the area 100, whether from the object 104 or otherwise, can be input by a microphone unit. The sensor unit can include a radar, an accelerometer (optional), a gyroscope (optional), an inertial measurement unit (optional), a photodetector (optional), a local temperature sensor (optional), a local humidity sensor (optional), and a particulate sensor (optional). The sensor unit is used for sensory capture of various inputs, as disclosed herein. For example, as disclosed herein, a sensory input identified in the area 100, whether from the object 104 or otherwise, can be input by a sensor unit. For example, a photodetector, a local temperature detector, a local humidity detector, or a particulate sensor can sense its surroundings and the data thereof for a processor to act or decide to act accordingly.For example, a photodetector, a local temperature detector, a local humidity detector, or a particulate sensor can sense fire, smoke, carbon monoxide, pollution, or other events such that the processor issues an alert via a communication unit (e.g., calling emergency services or a predetermined telephone number, turning off or deactivating a gas supply, electrical panel, or water main valve), a speaker unit (e.g., directing the object 104 out of the area 100), or a user interface unit. Thus, if the device 102 includes a communication unit and the processor is coupled (e.g., mechanically, electrically, logically) to the communication unit, the processor can take actions including instructing the communication unit to send a message to another device 102 (e.g., locally) or to a computer (e.g., desktop, laptop, cell phone) remote from the area 100, the message including content related to the object 104, the area 100, or the event.

[0068] FIG. 6 shows an internal view of FIG. 3 according to the present disclosure. FIG. 8 shows a photograph of the internal cavity of FIG. 3 based on FIG. 6 according to the present disclosure. In particular, device 102 includes a housing 202, a logic unit 204, a power supply unit 206, vents 208, a frame 210, a heat sink 212, a speaker 212, a window 214, a radar 216, and a microphone 218, some of which are shown in FIG. 5. Housing 202 hosts frame 210. Logic unit 204 is secured or mounted (e.g., fastened, mated, interlocked, glued) on frame 210 such that logic unit 204 extends between power supply unit 206 and window 214, although this positioning can vary. Power supply unit 206 provides power to logic unit 204. The power supply unit 206 is secured or attached (e.g., fastened, mated, interlocked, glued) to the housing 202 such that the logic unit 204 extends between the power supply unit 206 and the heat sink 212 or the radar or window 214. The power supply unit 206 faces the logic unit 204, although this positioning can vary. The air vent 208 is defined by the housing 202 such that the logic unit 204 extends between the air vent 208 and the window 214 of the speaker 212 (left side), although this position can vary. The air vent 208 is formed by a symmetrical set of openings, although this is not required and can vary (e.g., asymmetrical openings) if desired. The heat sink 212 is positioned between the radar 216 or window 214 and the logic unit 204 or the power supply unit 206, although this positioning can vary.

[0069] The vent 208 and heat sink 212 are used collectively as a group of passively cooled components, which may affect how the heat sink 212 is designed (e.g., size, shape) or positioned, or how the housing 202 is designed (e.g., size, shape), or how the vent 208 is designed (e.g., size, shape) or positioned, respectively, to optimize convection (e.g., direct rising heat away). This is technically advantageous because the vent 208 and heat sink 212 operate with minimal or no noise, which may be desirable in certain environments (e.g., elderly care recipients). However, a passively cooled group of components may be technically disadvantageous because the heat sink 212 may be larger than desired, may be economically expensive, may be space-constrained, or may not be readily available due to supply shortages. Similarly, passive cooling components may be technically disadvantageous because the amount of heat removed by the vent 208 or heat sink 212 is highly dependent on the ambient temperature of the air surrounding the heat sink 212 or housing 202 (e.g., next to a radiator, heater, or vent in winter), potentially causing the logic unit 204 or radar 216 to operate at a higher temperature. Therefore, in situations where passive cooling components are technically disadvantageous, active coolants may be used. The active coolant may be a cooling fan, a coolant circulation system, or other suitable coolant. When embodied as a cooling fan, the active coolant may forcibly disperse ambient air over the radar 216 or logic unit 204, providing more effective cooling than passive cooling components. However, active coolants (e.g., cooling fans) have moving or balancing parts and may generate noise due to the moving ambient air. Similarly, active coolants may also wear or rattle, which may be undesirable in certain environments (e.g., elderly care recipients). Similarly, active coolants can reduce the long-term reliability of the device 102, increase dust collection rates, increase the likelihood of mechanical failure or component imbalance, increase connector wear, and add noise to radar measurements, potentially requiring more filtering and correction thereof.Thus, depending on the use case, the radar may be actively or passively cooled.

[0070] 6, if the housing 202 includes two speakers 212, the logic unit 204 or heat sink 212 or radar 216 may extend between the speakers 212, although the speakers 212 may be located in other locations, whether inside or outside the housing 202. Note that there may be more than one speaker 212, one speaker 212, or no speaker 212 at all. The speakers are powered via the power supply unit 206 and controlled via the logic unit 204.

[0071] The radar 216 (e.g., a circuit board with a receiver and an antenna, or a transceiver with a transmit antenna and a receive antenna) extends between the window 214 and the heat sink 212, or the logic unit 204, or the power supply unit 206, although this location may vary. Because the radar 216 tracks (e.g., transmits and receives) through the window 214, there may be no window 214, but this minimizes blockage, interference, and noise. As shown in FIG. 6, the housing 202 includes two windows 214 that are symmetrical with respect to one another, although there may be more than two windows 214, or the windows 214 may be asymmetrical with respect to one another.

[0072] The microphone 218 is positioned adjacent to the window 214 so that the radar 216, heat sink 212, or logic unit 204 extends between the microphone 218 and the power supply unit 206, although the microphone 218 may be omitted or positioned elsewhere. The microphone 218 is powered through the power supply unit 206 vias and controlled by the logic unit 204 to supplement, enhance, augment the radar 216, or receive user input, as disclosed herein.

[0073] FIG. 10 illustrates an embodiment of a circuit board with an antenna set for a radar according to the present disclosure. FIG. 11 illustrates an embodiment of the scope of the device of FIG. 3 according to the present disclosure. In particular, as described herein, radar can operate on digitally coded time-of-flight principles, which can be embodied in various ways. Radar can be used to track or monitor the safety, health, welfare, or care of individuals managed by a processor. For example, radar can transmit a pulse and measure the time it takes for a reflection (e.g., an echo) to return from an object (e.g., a person, furniture, appliance, wall, floor, ceiling) (e.g., a receiver, transceiver, sensor, distance sensor). For example, radar can operate through a phased array of transmit and receive antennas to provide highly directional measurements. This is shown in FIG. 10, where the radar includes a circuit board with two antenna arrays (small gold rectangles, although other colors or shapes are possible) and a heat sink (small black rectangle, although other colors or shapes are possible) covering the main radio frequency processing chip. For example, the radar may include a set of phased arrays, each consisting of a set of patch antennas, allowing the radar to track objects 104 living or present within the area 100.

[0074] The device 102 includes a processor coupled (e.g., electrically, logically, mechanically) to the radar to achieve various design goals or capabilities. Some of these goals or capabilities may enable edge computing and hardware-accelerated artificial neural networks (ANNs), which may include various corresponding models to enable various raw sensing and processing capabilities for building various software-based applications supporting various types of use cases, as described herein. For example, some of the ANNs may include convolutional neural networks, recurrent neural networks, long short-term memory neural networks, or other suitable forms of ANNs. These forms of edge computing and hardware-accelerated ANNs may enable the processor to handle various use cases based on sets of data from the radar. These uses may include enabling the processor to track the heart rate or pulse of the object 104 based on sets of data from the radar by using various transmission characteristics of the radar, capture a portion of at least some of the “backscatter” (also known as the reflected signal) reflected by the object 104, and track the heart rate or pulse of the object 104. These use cases may include the processor tracking breathing rate based on a set of data from a radar, which may be tracked similarly to a heart rate, as described herein, or processing a signal to identify certain slow respiratory fluctuations based on a set of data from a radar. These use cases may include the processor tracking the location of objects 104 within an area 100 based on a set of data from a radar, i.e., permanent knowledge of the location of each individual moving within the area 100 (within the field of view 130). These use cases may include the processor tracking fall detection based on a set of data from a radar, i.e., the ability to detect the orientation of an object 104 and its rate of change.Such use cases may include the processor being programmed for a particular room geometry based on a set of data from the radar, i.e., the location of large planes (e.g., floor, walls, ceiling, furniture, appliances). These use cases may include the processor tracking the pose of the object 104 based on a set of data from the radar (the orientation of the object 104). These use cases may include the processor performing object classification based on a set of data from the radar. Identification of furniture or appliances can provide context for whether someone is sleeping in a bed or lying on a table. These use cases may include the processor tracking various activities of daily living based on a set of data from the radar, i.e., the ability to accurately determine the timing and frequency of life patterns including microbehaviors such as sleeping, drinking, eating, toileting, socializing, and taking medication.

[0075] Given the ongoing chip shortage, device 102 may be designed to leverage 5G chips to reduce or mitigate this risk. For example, radars are all-digital designs (non-all-digital or hybrid designs are also possible), allowing them to leverage existing communications chips (e.g., wireless RF receivers, wireless RF transmitters, wireless radio frequency transceivers, wired interface cards, and waveguides). For example, 5G (or 3G, 4G, or 6G) cellular technologies may operate in the same or similar bands as radars operating in the Ku, K, and Ka bands, leading to significant and ongoing investments in components for wireless systems in these frequency bands. Similarly, further performance improvements driven by competition in the cellular market may impact radar. In contrast, some radars rely on complex analog designs, and only a small proportion will benefit from these external market developments.

[0076] As described herein, the device 102 can be used to track where the object 104 is located within the area 100 based on a set of data received by the processor from the radar and processed by the processor, and to detect a fall of the object 104 within the area 100 based on a set of data received by the processor from the radar and processed by the processor. For example, if the object 104 is a human, the radar can detect the human's position and orientation through at least one or two standard US-style walls (e.g., drywall), and the radar can "see" adjacent rooms, bedrooms, bathrooms, hallways, pantries, or other areas 100 within its field of view 130, as shown in FIG. 11 . However, because the radar may have coverage gaps (e.g., occlusion), the device 102 may be supplemented by another device 102 (a "satellite" unit) located within or tracking the area 100, or by sensors (e.g., cameras, microphones, motion sensors, proximity sensors, distance sensors) located within or tracking the area 100. The devices 102 (e.g., housings) or sensors can be spaced apart to increase resolution or coverage of the object 104 or area 100. This allows the device 102 to validate, corroborate, or confirm its data, analysis, or conclusions, or to infer something about the object 102 or area 100 that the device 102 cannot track or identify. For example, if the object 104 is a human, the bathroom is a common place in the home for falls. The radar may generate a set of data with a slightly reduced quality or resolution while the object 104 is showering, but this does not prevent the detection of a fall. Some parts of the bathtub may cast a shadow on the sensor, but the radar will detect the fall because this occurs when the radar's view of the object 104 is unobstructed or unblocked.However, if the radar view of the object 104 is obstructed or occluded, the device 102 can communicate (e.g., wired, wireless, waveguide) with another device 102 located within or capable of tracking the area 100, or sensors located within or capable of tracking the area 100, to individually or collectively determine whether the object 104 is experiencing an event, such as a fall, as disclosed herein. For example, the second device 102 can be tracked at Ku-band, K-band, or Ka-band, as disclosed herein. This configuration thus provides complementary tracking or sensing capabilities that can include using low-resolution infrared sensing (or other sensing modalities, including ultrasonic sensors, LIDAR, radar, motion sensors, proximity sensors, distance sensors, or others, whether low-resolution or high-resolution, whether line-of-sight or non-line-of-sight). "Satellite" devices 102 can be deployed in distant rooms to provide coarse or fine sensing of people's locations within the area 100, extending coverage to adjacent spaces that may be blocked (e.g., by metal walls). "Satellite" devices 102 can also have a microphone or set of microphones, which can help detect specific falls and provide two-way communication (e.g., via a Bluetooth unit, Wi-Fi unit, RJ). Some, many, most, or all of the computations are performed on the main unit 102, although these computations may be distributed or shared with other devices 102 (e.g., via a communications unit). Additionally, multiple main or "satellite" units can be deployed to extend the tracking range and provide "rear" coverage.

[0077] Thus, if a device 102 is a first device 102 (having a first radar that transmits a first set of first tracking data to a first processor) positioned within the area 100 to track a residence or object 104 located therein, there can be a second device 102 (having a second radar or sensor that transmits a second set of second tracking data to a second processor) positioned within the area 100 to track or sense a residence or object 104 located therein. The first device 102 and the second device 102 may be spaced apart from each other. For example, the first device and the second device may be opposite each other. Thus, the second device 102 can operate within the area 100 without interfering with the first device 102 operating within the area 100 (e.g., radar interference) when a second radar (e.g., time-of-flight) or sensor (e.g., sensor unit, audio unit, vision unit) operating within the area 100 is shielded or unshielded from tracking the object 104 living or located within the area 100. The second radar or sensor generates a second set of data based on tracking or sensing the object 104 living or located within the area 100 when the first radar operating within the area 100 is shielded or unshielded from tracking the object 104 living or located within the area 100. The second radar or sensor transmits a second set of data to a second processor when the first processor operating within the area 100 is shielded or unshielded from tracking the object 104 residing or located within the area 100, such that the second processor determines, based on the second set of data, whether the object 104 is experiencing a second event within the area 100 and takes a second action in response to the second event determined to be occurring within the area 100.The first and second radars can both operate in the same band (e.g., K band, Ku band, Ka band) or the same range within the same band (e.g., between about 23 GHz and about 25 GHz, inclusive). The first event and the second event may or may not be the same event, and the first action and the second action may or may not be the same action. The first and second radars or sensors may or may not have overlapping fields of view. The first and second devices 102 can be configured to enable a second processor to communicate with the first processor (e.g., via a communication unit, a Bluetooth unit, a Wi-Fi unit), thereby enabling the first processor to confirm or validate the first set of data based on the second set of data, determine whether the object 104 is experiencing an event in the area 100 based on the second set of data that confirms or validates the first set of data, and take action in response to the event determined to be occurring in the area 100. The second processor may be configured to communicate directly (eg, housed in a common housing or having a common controller) or indirectly with the first processor.

[0078] As described herein, the radar may operate in the Ku, K, or Ka bands and track objects 104 within the area 100 by emitting waves. These waves are a type of non-ionizing radiation, and exposure to high-frequency non-ionizing radio waves is regulated by the Federal Communications Commission and other international regulatory agencies by limiting the "dose" in terms of the amount of energy absorbed by the body (e.g., object 104). Thus, in some embodiments, the radar's transmit power is programmable, and the radar can be configured to comply with or conform to all appropriate regulations in accordance with guidance for uncontrolled exposure to the general public (i.e., everyday use), which can occur via over-the-air updates.

[0079] As described herein, radar can sense through typical walls found in a typical home (e.g., wood or metal frame with drywall and siding, or wood or metal frame with a pair of drywall). The radar can be a time-of-flight radar, a Doppler radar, or other suitable type of radar. Time-of-flight radar can measure the time it takes a pulse to travel to a reflecting object and back (e.g., an echo) or allow for direct observation of stationary objects. For example, typical household items (e.g., furniture, housing) are stationary, and if the object 104 is an elderly or frail human, the object 104 may be similarly slow-moving. Doppler radar can measure the movement of a reflecting object toward or away from the Doppler radar. When using Doppler radar, the faster the object moves, the more of it is visible (e.g., more data is obtained). Generally, human observation relies on the argument that humans are never stationary, and therefore something about their presence is always detectable. However, this means that flapping fabric will be detected and may be distracting. Stationary objects may be missed. Of these two radar modalities, time-of-flight radar is far more suited to observing relatively slow-moving domestic environments. Naturally, time-of-flight systems can observe motion by creating "video" from still frames. As disclosed herein, hybrid designs are also possible (e.g., radar mode switching) that use Doppler radar to augment, supplement, substitute, strengthen, or replace time-of-flight radar in detecting tremors, skin movement due to heartbeat, chest flexion due to breathing, or other vibratory motion. Generally, such vibratory motion is slow relative to the radar wavelength, so the Doppler shift is often small and difficult to detect, although it is possible. However, higher frequency harmonics resulting from vibratory motion can be more easily detected. For example, if the object 104 is a human, a heart beating at 60 beats per minute can generate characteristic harmonic vibrations of several hundred Hz.Doppler radar can detect these harmonics and allow the device 102 (e.g., a processor) to distinguish some or various vibrational motions from intentional movement. For example, this disambiguation can manifest itself in detecting respiration, where a natural human breathing rate is on the order of a few Hz. However, human movements such as swinging your arms while walking occur at a similar frequency. By detecting the high-frequency harmonics caused by breathing rather than swinging your arms, Doppler radar gains the technical advantage of being able to detect respiration rate despite movement. It should be noted that Doppler radar (and Doppler information), as disclosed herein, can be used for people monitoring. It should be noted that some versions of existing Doppler radar products tend to have fairly limited performance in people monitoring tasks. These products may lose the ability to detect activity within about 5 feet (about 1.5 meters) of the source. Similarly, measuring heart rate and respiration rate requires standing still and remaining about 1 to 2 feet (about 1.5 m to 2.5 m). Time-of-flight radar, Doppler radar, or any other suitable radar may be used.

[0080] As described herein, device 102 can be supplemented, enhanced, or augmented with data from line-of-sight sensing (e.g., camera, LIDAR, high-frequency radar). For example, many technologies rely on a direct line-of-sight between the sensor and the object 104 being detected. Some optical camera-based, infrared-based, acoustic-based, or LIDAR-based solutions rely on line-of-sight. Some high-frequency radar (specifically, some versions at about 60 GHz or about 76 GHz to about 81 GHz) has very poor material penetration and is similarly substantially limited to line-of-sight, but can be used to supplement device 102. Some line-of-sight-based systems have difficulty uniformly covering a typical indoor environment. Rooms are not rectangular, and furniture creates occlusions, presenting significant challenges. For example, a person who falls behind a table or counter is likely not visible to a single indoor sensor. In contrast, the radar disclosed herein can directly track a person behind an obstacle. Furthermore, some line-of-sight technologies only operate in a single room, requiring the installation of a unit in each room and residential space to provide sufficient coverage. In addition to the various financial costs of these devices, such setups also require professional installation, making the business model unviable. In contrast, the device 102 can cover residential spaces up to approximately 2,000 square feet (although potentially more or even more, depending on various factors), can operate through walls, and can be expanded with additional base units (e.g., another device 102 or sensor) for wider coverage (e.g., larger homes, multiple floors), or can use “satellite” devices 102 to provide coverage for enclosed adjacent spaces. Furthermore, computer vision-based systems require people to accept the placement of cameras in their most private spaces, which is difficult to achieve (e.g., user surveys in the United States have shown that fewer than 42% of older adults would accept a camera in their bedroom, and even lower in Europe).

[0081] FIG. 12 illustrates an embodiment of a set of microphones for the device of FIG. 7 according to the present disclosure. FIG. 13 illustrates an embodiment of a microphone for the device of FIG. 7 according to the present disclosure. FIG. 14 illustrates an embodiment of raw readings from the device of FIGS. 1-7 and a virtual skeleton formed by the device of FIGS. 1-7 from the raw readings according to the present disclosure. In particular, the device 102, as shown in FIG. 5, can include a microphone unit and be controlled by a processor for various purposes. For example, one of these purposes is to supplement (e.g., validate, verify, corroborate) the data set generated by the radar and sent to the processor for processing, as disclosed herein. For example, the microphone unit can be used to receive voice commands from the object 104 to control a voice assistant (e.g., Siri, Google Assistant) running on the processor. The microphone unit can include a microphone or set of microphones. The microphone unit is sensitive and robustly constructed to enable accurate and satisfactory use, be expandable to support increasing functionality over time, and be cost-effective. The microphone unit provides a "voice-first" conversational user experience, enables multi-party communication (e.g., from device 102 to device 102 or from device 102 to a remote phone or computer), and, with user consent, provides rich context about people and their activities. In such situations, device 102 includes a corresponding speaker unit controlled by the processor.

[0082] The device 102 uses a microphone unit and a speaker unit to converse with the subject 104 (or its agent) for easy setup without the need for the device 102 to download a dedicated app, respond to conversational voice commands, enable communication with loved ones, caregivers, and caregivers (e.g., pre-programmed call numbers, chat names), and call for help from the subject 104 when the device 102 detects imminent danger or when the subject 102 calls for help. Note, however, that other embodiments may include apps downloaded from an app store (e.g., Google Play, iTunes).

[0083] The microphone unit is used to capture sounds indicative of a fall, as determined by the processor, which can help confirm or validate sounds detected by the radar and reduce false positives. For example, if data, information, or instructions from the radar conflict with the microphone unit, the radar can take control, as determined by the processor (however, this can be reversed, or the control default is not currently selected or available). The microphone unit can capture loud sounds, such as a crash or breaking glass, which may indicate other safety-critical situations, intrusions, fires, or weather events, as determined by the processor. The microphone unit can capture sounds from the kitchen and bathroom that are indicative of various daily activities and self-care, as determined by the processor. The microphone unit can capture sounds that may indicate social activities or media consumption, as determined by the processor. The microphone unit can capture airway sounds, such as coughing, wheezing, difficulty breathing, and snoring, which can indicate whether a lung disease is beginning or worsening, as determined by the processor. The microphone unit can capture tones of voice indicative of emotions (sadness, anger, frustration, happiness) or increased depression, as determined by the processor. The microphone unit can learn from the subject 104 over time and capture additional audio context associated with specific behaviors or symptoms of the subject 104.

[0084] The microphone unit can include an array of three (but two or more) digital microphones (although analog microphones are also possible). These microphones have a good signal-to-noise ratio and help the device 102 hear commands and other sounds from as far away as possible (e.g., within about 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, or 1 feet or meters of the room). In addition to these microphones, the microphone unit subsystem can include a signal processing chip that provides 3D direction-of-arrival information for sound sourcing.

[0085] The microphone unit can include two different microphone processing chains, either of which can be performed by the microphone unit's processor. The first processor performs echo cancellation and sound separation in preparation for processing by the voice assistant. The second processor is a direction of arrival analysis processor that uses all three microphones to separate multiple sound sources and locate where these sounds originate within the monitored area 100. Thus, the processor can localize sound sources (e.g., voices, falling sounds) so that the processor (e.g., its edge AI algorithms) can correlate these sources or sounds with radar data or visual data during training. Similarly, the processor can use this sound information as location information to augment activity recognition. Similarly, the processor can also separate distinct sounds. For example, isolating background noise from a television or vacuum cleaner from human sounds can improve the processor's ability to hear commands and recognize sounds. Thus, if device 102 includes a set of microphones and a processor coupled to and controlling (e.g., receiving data from) the set of microphones, the processor can be programmed to activate the set of microphones so that the set of microphones receive a set of acoustic inputs generated from a residence or object 104 located within area 100, the processor isolates the set of acoustic inputs, transmits the set of acoustic inputs to the processor based on the isolation of the set of acoustic inputs, locate where within area 100 the set of acoustic inputs originated, confirm or validate a set of data based on the identification of where within area 100 the set of acoustic inputs originated, and take action responsive to an event determined to be occurring within area 100 based on the confirmation or validation of the set of data by the set of acoustic inputs. Note that while the microphone units may have their circuit boards mounted on top, as shown in FIG. 12 , this is a requirement and the circuit board may be located elsewhere.

[0086] As described herein, device 102 may include a vision unit (e.g., optical camera, infrared camera) controlled by a processor for various purposes. For example, one of these purposes is to supplement (e.g., validate, verify, corroborate) the set of data generated by radar and sent to the processor for processing, as disclosed herein. For example, the vision unit may be used to receive gesture commands from subject 104 to supplement the control of a voice assistant running on the processor. For example, the vision unit may include a high-resolution (e.g., 4K, 8K) video camera (e.g., red, green, blue) that the processor can optionally enable as needed, and this may be enabled with user consent. This video camera may operate in various lighting conditions and augment the various sets of data available to the processor to enhance the training of various edge AI algorithms, particularly for identifying people via the radar subsystem. While device 102 may include the resulting edge AI algorithms, this is not required, and subject 104 (or its agent) may opt out of or disable the vision unit as desired.

[0087] The video camera may use a wide-angle lens (or even a fisheye lens) to ensure a good view of the room in which it is installed. Such a camera may be an integrated module containing a sensor, a lens, and supporting electronics. Note that imaging data may be preprocessed (e.g., dewarped) before image processing. The video camera may provide imaging data to a processor to complement the set of data the processor receives from the radar. The imaging data may be used to support training and validation of the location, posture, or identity of a person within the video camera's field of view. The processor may compare this imaging data with a set of data received from the radar or data received from an audio unit, at least within the area 100 in which the device 100 is located. Furthermore, imaging data may be useful for recognizing daily activities and training radar before it is used by the object 104 or while the radar is tracking the object 104. There are also techniques for directly extracting vital signs, such as heart rate, from the imaging data, which may be integrated for training various edge AI algorithms. Because the vision unit may be used in a variety of conditions (e.g., daytime, nighttime), the video camera may operate in lighting conditions ranging from full sunlight to complete darkness. To meet this need with a single camera (although multiple cameras could be used, feeding into a processor), the video camera may use a dual-band filter capable of observing both the visible spectrum (red, green, and blue) and near-infrared light. The device 102 may also include its own near-infrared light source for nighttime operation. While there may be drawbacks to using this dual-band filter (e.g., mild distortion of red-green-blue colors), as disclosed herein, this may be compensated for and may not affect various applications.Thus, if device 102 includes a camera (e.g., optical or infrared) and a processor is coupled (e.g., mechanically, electrically, logically) to the camera for controlling the camera (e.g., the processor can be programmed (or a user can instruct) to activate the camera to receive images generated by live objects 104 within area 100, the processor confirming or validating a set of data, and take action responsive to an event determined to be occurring within area 200 based on the confirmation or validation of the set of data). The camera can include a dual-band filter configured to enable observation in the visible spectrum (RGB) and near-infrared light.

[0088] As described herein, the device 102 includes a processor (e.g., a controller, an edge processor, a single-core processor, a multi-core processor, a system-on-chip, a graphics processing unit, a hardware accelerator, a neural network accelerator, a machine learning accelerator) and a radar (e.g., a time-of-flight radar, a Doppler radar), where the processor is coupled (e.g., electrically, logically, mechanically) to the radar to control the radar (e.g., receive tracking data). For example, the processor can enable local or edge computing to improve processing speed or provide data privacy or data security. For example, sensors (radar, acoustic, visual, etc.) provide data (radar, acoustic, visual, etc.) to an edge processor. This is technically advantageous for several reasons. First, it allows for respect for people's privacy because some, much, most, or all of the collected data (radar, acoustic, visual, etc.) rarely or never leaves the device 102 (as is possible with cloud computing). Second, it minimizes bandwidth usage and power consumption, reducing costs. Third, on-board processing minimizes latency, improving the user experience and responsiveness of the sensor platform, and identifying events as they occur.

[0089] This processor, which realizes an edge computing platform, can include a single-core processor or a multi-core processor, regardless of whether these cores are local or remote from each other within the monitored physical area. A multi-core processor can include multiple independent cores. For example, a multi-core processor is a computing component with two or more independent processing units that read and execute program instructions, such as front-end applications, through multiprocessing or multithreading. Program instructions are processing instructions such as addition, data movement, or branching. A core can execute multiple instructions simultaneously, thereby improving the overall operating speed of front-end applications suitable for parallel computing. As disclosed herein, the cores can operate in parallel while adhering to the principles of atomicity, consistency, isolation, and durability (ACID) when accessing files or other data structures simultaneously, ensuring that operations / transactions (e.g., read, write, erase) on such data structures are processed reliably. For example, a data structure can be accessed simultaneously through at least two cores without locking the data structure. For example, as disclosed herein, figures and text can be processed simultaneously. Note that there can be two, three, four, five, six, seven, eight, nine, ten, twelve, tens, hundreds, thousands, millions, or more than two cores. Cores may or may not share caches, and cores may or may not implement inter-core communication via message passing or shared memory. Common network topologies for interconnecting cores include bus, ring, two-dimensional mesh, and crossbar. Homogeneous multicore systems contain only identical cores, while heterogeneous multicore systems can have non-identical cores. Cores in a multicore system can implement architectures such as very long instruction word (VLIW), superscalar, vector, or multithreading.Additionally or alternatively, an edge computing platform, whether a stationary or mobile platform, may include a graphics card, a graphics processing unit (GPU), a programming logic controller (PLC), a tensor core unit, a tensor processing unit (TPU), an application specific integrated circuit (ASIC), or another processing circuit.

[0090] This processor can use the latest generation of edge processing chips to power advanced AI at the edge while balancing cost and performance. As shown in Figure 5, the processor can include both a powerful central processing unit (or another form of processing logic unit) and a discrete ANN accelerator, enabling diverse workloads. The processor can be paired with sufficient memory (RAM) and persistent storage to provide a balanced platform. For example, there can be 4, 6, 8, 12, 16, 32, 64, 128 gigabytes of RAM or persistent memory (e.g., flash memory), or two or more memories.

[0091] As also shown in FIG. 5 , device 102 can include a communications unit for partial or complete connectivity options (e.g., wired, wireless, waveguide). Internet (or LAN or WAN or network) connectivity can be provided via either a Wi-Fi or cellular modem or other suitable communications device. The communications unit enables device 102 to function as a hub for multiple devices connected via either Bluetooth (e.g., a low-power or other suitable personal network) or Wi-Fi (or Li-Fi or other suitable connection). To keep device 102 up to date, the communications unit supports over-the-air updates via either a Wi-Fi or cellular data connection (or other suitable connection). Similar to how Tesla updates its cars, device 102 is configured to support evolving and expanding use cases, and as various edge AI algorithms and software evolve, the capabilities latent in each version or device 102 are unlocked and delivered transparently to the user. For example, this radar can cover a typical approximately 2000 square foot (or more) home with standard U.S. drywall construction, detect falls indoors and through walls, detect breathing and heart rates in static rooms up to approximately 20 feet (or more) away, and continuously detect an individual's pose, such as sitting, standing, or lying down. Similarly, device 102 can include wearable biosensor patches (e.g., on arms, torso, neck, or legs) to provide tagged data (e.g., wireless communication) for further training and validation of various edge AI algorithms for heart rate, breathing, posture, and fall detection. Similarly, device 102 can include training AI to account for encountering various home configurations that may not have been anticipated during testing (e.g., for future proofing against unforeseen scenarios)—this training can continue indefinitely into the future.

[0092] As described herein, the processor can use geofencing for a variety of purposes. Because radar can see through walls, it may view spaces it should not, such as a neighbor's property. To prevent this, the processor can be programmed to allow a user to geofence an area of interest or define an area of interest for which data will be ignored. The object 104 (or its agent or user) can define the boundaries of the geofence. One way to do this is to notify the radar when the user walks within the perimeter of the area 100, indicating to the radar when it should take measurements of the user's location. Notification can occur via a computing device (e.g., a handheld unit, a wearable unit) with a button that the computing unit communicates (e.g., wired, wireless, or waveguide) with the device 102 (e.g., a clicker), a mobile app (e.g., via a cell phone or tablet) that communicates with the radar, a sound picked up by a microphone unit, or a gesture by the user captured by the vision unit. For example, the device 102 may include a clicker (e.g., handheld unit, wearable unit) that communicates (e.g., wired, wireless, or waveguide pairing) with the device 102 (e.g., via a communication unit, Bluetooth unit, Wi-Fi unit), or a mobile application executable on a mobile device (e.g., mobile phone, tablet computer) that communicates (e.g., wired, wireless, or waveguide pairing) with a processor on the mobile device (e.g., via a communication unit, Bluetooth unit, Wi-Fi unit).In this manner, the user may be instructed, or the processor may be programmed, to define an area based on (a) moving the clicker within or outside area 100 (e.g., walking around area 100) and activating the clicker within or outside area 100 (e.g., stopping and activating, or walking and activating, or walking while active), or (b) interacting with a mobile application (e.g., mobile phone, tablet computer) located within or outside area 100 before taking action in response to an event determined by the processor to be occurring within area 100.

[0093] Another method of geofencing is to record the shape and identification of rooms. Through a mobile application running on a mobile phone or tablet computer that communicates with the device 102's communications unit, various rooms / locations within the area 100 are indicated as part of setup or re-setup, calibration or re-calibration, or onboarding or re-onboarding. The object 104 can be guided by the mobile app to walk around by pressing buttons on the mobile phone or tablet computer, marking the area 100. Note that a clicker may additionally or alternatively be used, and the setup or re-setup, calibration or re-calibration, or onboarding or re-onboarding may be guided by the device 102's user interface or speaker unit. Furthermore, similar approaches (e.g., clicker, walking by pressing buttons on the mobile phone or tablet, hand signals captured by a vision system, or voice commands captured by a microphone unit) can also be used to mark boundaries for geofencing or blocking from surveillance or observation. The marked area can be displayed as a room shape or floor plan from a radar perspective, indicating the field of view. Additionally, there may be installer-based setups that include beacon devices (e.g., wireless beacons) to better track room boundaries and fences. The device 102 may support a setup mode in which the device 102 (e.g., a communication unit) listens for beacon transmissions of a code that matches the setup during the setup mode for better fencing accuracy.Thus, the device 102 includes a beacon (e.g., a housing) having a transmitter (e.g., radio) configured to transmit a signal (e.g., wireless), the device 102 includes a receiver (e.g., radio) configured to receive the signal, and a processor is coupled (e.g., mechanically, electrically, logically) to the receiver, such that the radar and the beacon are spaced apart from each other (e.g., within about 5, 4, 3, 2, 1 feet or meters), the transmitter can transmit the signal, the receiver can receive the signal, and the processor can determine whether the object 104 is experiencing an event within the area 100 based on the set of data and the signal, and take action in response to the event determined to be occurring within the area 100.

[0094] Another way to do this may be during the setup process, when the processor is placed in geofencing mode, and the processor may assume that the user is only moving within the desired recording area 100. The user then moves freely within the area 100 tracked by the radar. The processor records the extreme edges or corners of this movement (e.g., voxel-based processing). When the movement is over, the user returns to the device 102 and notifies the device 102 that the movement is over. The processor then performs additional processing on the data to determine the extremities of the geofence. Another way may be when the user walks around the area 100 with a mobile app (e.g., phone, tablet, wearable) that instructs the user (e.g., via a speaker, display, or via) how the user should walk around the area 100 when the user is in and out of range of the radar, based on the mobile app communicating with the processor via the communication unit. This allows the radar to use the user as an active probe, and the mobile app can notify the user (e.g., via speaker, display, or vibrator) if the user is in the shadow of the sensor or out of range of the radar (such that movement within the area 100 is not detected), or to remain in a location within the area 100 longer so that the radar can collect additional measurements to reduce sensing noise. Alternatively or additionally, the user can walk inside or outside the area 100 where the radar should not see. This could be outside the area 100 (e.g., a residence, a house), as shown in Figures 1-2, and could help reinforce the boundaries where the radar should see, or could help define specific areas where the user does not want the radar to see (i.e., for privacy reasons). This allows the user to define a geofence-defined area with holes (i.e., a "polygon with holes").Alternatively or additionally, the user may be instructed to carry a large reflective object (e.g., metal object, mirror) while walking around the area 100 to simplify the radar's tracking of the user. Furthermore, the user may be instructed to input the height of the radar above the floor and below the ceiling within the area 100, or the processor may assume these heights as a default. The user may also provide or draw a map of the area 100 on a computing device (e.g., a mobile phone or tablet touchscreen) and place the radar within the map for access by the processor of the device 102. Thus, the user may be instructed, or the processor may be programmed, to define the area 100 before the processor takes action in response to an event determined to be occurring within the area 100. This may occur based on the user moving within the area 100 before the processor takes action in response to an event determined to be occurring within the area 100. This may occur based on the user moving outside the area 100 before the processor takes action in response to an event determined to be occurring within the area 100. This may occur based on the user hosting a reflector trackable by radar before the processor takes action in response to an event determined to be occurring in area 100. This may occur based on the processor accessing a value corresponding to a height above floor before the processor takes action in response to an event determined to be occurring in area 100. This may occur based on the processor accessing a map of area 100 before the processor takes action in response to an event determined to be occurring in area 100. The user may instruct the processor to create a map of area 100 (e.g., on a personal computing device in communication with the communication unit of device 102) before the processor takes action in response to an event determined to be occurring in area 100.

[0095] As described herein, the processor can ignore portions of the area 100 when processing a set of data received from the radar. This can occur for each beam position (also known as a steer) where the radar scans the field of view in a "lawnmowing" pattern. Each position in the scan is called a steer. Each position in the scan returns a list of numbers, with each element representing a reflector at a certain distance from the radar. The result of this scanning process is a 3D voxel-based map of space formed by the processor. Voxels in this map can be zeroed out if they are outside the geofenced area. If all voxels within a steer should be zeroed out, the processor can adjust the scan pattern to avoid certain steers to speed up detection of other scan areas. In background subtraction, once the 3D voxel map arrives at the processor, it can be compared to a voxel map representing areas the radar should not record. Voxels that should not record can be zeroed out. After location detection, the set of data from the radar is processed by the processor to calculate the location of the person / object. If the location of the person / object is outside the area 100 (e.g., a polygonal area), no further processing is performed on the person / object and the location is not reported / stored. After machine learning, the set of data from the radar is processed by the processor to calculate virtual skeleton keypoints of the object 104 living or located within the area 100, as formed by the processor based on the set of data from the radar. Depending on a threshold of how many virtual skeleton keypoints are outside the allowed sensing area, the virtual skeleton keypoints may not be reported, as shown in FIG. 14. In this manner, the processor can form a three-dimensional map of the area 100 based on the set of data based on the radar and scans from within the area 100 that are outside the area 100, such that the three-dimensional map has zeroed areas before the processor takes action in response to an event determined to be occurring within the area 100.The defined area may include a first volume of space and a second volume of space, and the processor may be programmed to access a threshold associated with the first volume and request the radar to adjust based on a zeroed region meeting the threshold to not track the first volume to facilitate the radar tracking the second volume before the processor takes action in response to an event determined to be occurring within the area 100. The processor may access a scan map of the area 100 formed based on the set of data and having a set of voxels, access a no-scan map of the area 100 for the radar, compare the scan map to the no-scan map, identify a subset of voxels from the set of voxels, and zero out the subset of voxels before the processor takes action in response to an event determined to be occurring within the area 100. A radar operating within the area 100 may track an object 104 outside the area 100, generate another set of data based on tracking the object 104 outside the area 104, and send the another set of data to a processor so that the processor can determine whether the object 104 is outside the area 100 and take another action (e.g., discard, remove, delete, or ignore the another set of data) in response to determining that the object 104 is outside the area 100. For example, as shown in FIG. 14 , the processor can form a three-dimensional skeletal model simulating the object 104 based on the set of data, determine whether the object 104 is experiencing an event within the area 100 based on whether the three-dimensional skeletal model meets or does not meet a threshold (e.g., matches a virtual skeletal signature for a given event), and take an action responsive to the event determined to be occurring within the area 100 based on whether the three-dimensional skeletal model meets or does not meet the threshold.For example, the processor forms a three-dimensional area model simulating the area 100 based on the set of data and a three-dimensional skeletal model simulating the object 104 within the three-dimensional area model based on the set of data, determines whether the object 104 is experiencing an event within the area 100 based on whether the three-dimensional skeletal model within the three-dimensional area model meets or does not meet a threshold (e.g., matches a virtual skeletal signature within the virtual model area to a predetermined event), and takes action in response to an event determined to be occurring within the area 100 based on whether the three-dimensional skeletal model within the three-dimensional area model meets or does not meet a threshold. The processor determines whether the three-dimensional skeletal model (with or without a virtual model region) satisfies or does not satisfy a threshold based on a set of virtual movements of the three-dimensional skeletal model (e.g., joint movements, torso movements, head movements, arm movements, leg movements, neck movements, end effector movements), identifies a set of atomic movements of the three-dimensional skeletal model corresponding to the set of virtual movements (e.g., joint bending and straightening, elbow bending and straightening, leg bending and straightening, torso movement), and correlates the set of atomic movements to an event.

[0096] As described herein, the processor can reset the geofence based on radar movement. The radar can detect its own movement in conjunction with an accelerometer, gyro, inertial measurement unit, or geolocation unit (e.g., GPS, GLONASS) and prompt the user (e.g., via a mobile app communicating with a speaker unit, user interface unit, or communications unit) to restart the geofencing routine. Alternatively, the device 102 can use its sensors to determine how the radar moved and update its internal representation of the geofence. The processor can save a master voxel map representing the space (without human / object movement) during the previous geofence measurement. This can be done by averaging voxel maps recorded during the previous geofence or by explicitly detecting an object 104 and subtracting its effect from the voxel map. The processor then calculates the same voxel map at the new location and calculates a coordinate transformer to realign the master voxel map to the new voxel map. The learned transformer can be used to transform the original geofence information into the processor's new coordinate frame. In this manner, the processor may access the movement threshold before taking an action in response to an event determined to be occurring within the area 100, access a geofence created by a user (e.g., via a user interface unit or a mobile app in communication with the communication unit of the device 102) before taking an action in response to an event determined to be occurring within the area 100, and take another action, including the geofence, based on the movement threshold being met before taking an action in response to an event determined to be occurring within the area 100. The other action may include modifying the geofence, resetting the geofence to a default state, or initiating a user guide (e.g., via the user interface unit or speaker unit, or via a mobile app in communication with the communication unit of the device 102) to re-geofence.The other action may include determining how the radar moved within the area 100 before the processor takes action in response to an event determined to be occurring within the area 100, and modifying the geofence based on how the radar moved within the area 100 before the processor takes action in response to an event determined to be occurring within the area 100.

[0097] As described herein, the processor can adjust the geofence due to changes in the environment (e.g., the purchase or construction of a new space, as shown in FIGS. 1-2 , or permission from an adjacent owner to view into their space, or the reconsideration of a previous space owned by the user but removed from view due to another concern, such as personal privacy or visitors). This can happen in a variety of ways. The user can present (e.g., via a user interface unit or mobile app in communication with the communication unit of device 102) a map of their space formed based on a set of data from the radar. The user can then remove or add spaces (e.g., via a user interface unit or mobile app in communication with the communication unit of device 102). The user can repeat the geofencing procedure described above. The user can notify the processor (e.g., via a user interface unit or mobile app in communication with the communication unit of device 102) whether the user wants to add or remove a space, and the geofencing procedure outlined above can be repeated for only the new spaces the user wants to add or remove. In this manner, the processor can access the geofence before the processor takes action in response to an event determined to have occurred within the area 100, and modify the geofence before the processor takes action in response to an event determined to have occurred within the area 100 and in response to the processor accessing a user input (e.g., a user interface unit or a microphone unit) indicating that the processor should modify the geofence.

[0098] As described herein, the processor functions to recognize toileting, dressing, eating, or other activities of daily living. Recognizing these activities is easy if the radar knows how the space is designed. For example, handwashing is likely to occur in the bathroom, while cooking is likely to occur in the kitchen. Or, eating may occur in the kitchen or dining room, but is unlikely to occur in the bedroom. Thus, the processor may be programmed to identify functional spaces within the area 100. Identifying rooms can proceed similarly to setting the room geometry or defining the area 100 (e.g., geofencing). The user then explicitly labels each identified space (e.g., via a user interface unit or a mobile app communicating with the communication unit of the device 102). Labeling can be performed by an app via a mobile phone or tablet operated by the user, where the user speaks the name of the room and a microphone unit records the utterance, or by the user drawing and defining the space while looking at a map generated by the processor (e.g., via a user interface unit or a mobile app communicating with the communication unit of the device 102). The processor can infer the use of a space from multiple probabilistic priors (statistics learned from a general population). For example, a person who remains relatively motionless in a horizontal pose for several hours in the evening is likely to be sleeping, as inferred by the processor. The surrounding area is likely to be a bedroom, as inferred by the processor. As a more complex example, a certain type of radar return corresponds to a person sitting, as inferred by the processor. The space where the person usually sits, as inferred by the processor, could be a bathroom, a favorite chair, or a dining table. If the person only sits in that location for a short time, as determined by the processor, it is likely that the person is using the bathroom, as inferred by the processor, and the area immediately surrounding that activity is likely to be a bathroom. Note that space need not be defined by walls, steps, or straight lines. Rather, space could be a more general concept of probability density.That is, certain activities are more likely to occur in certain areas of the house, as inferred by the processor. Just as activities can be used to identify or infer the space type inferred by the processor, knowing the common name of a space can be used as prior information for the processor to identify activities. If a space is known by the processor to be a kitchen, it is less likely that someone is sleeping or using the toilet in that space, as inferred by the processor, and more likely that someone is preparing a meal, opening the refrigerator, or eating, as inferred by the processor. This type of inference can be solved with a Bayesian network. In this manner, a user may be instructed, or a processor may be programmed, to assign identifiers (e.g., kitchen, bathroom) to sub-areas within area 100, such that the processor determines, based on the set of data and the identifier, whether object 104 is experiencing an event in the sub-area and take action in response to the event determined to be occurring in the sub-area. The identifiers can be assigned to sub-areas via a user operating a communication unit (e.g., via a communication unit, Bluetooth unit, Wi-Fi unit) with the device 102 such that the processor determines, based on the set of data and the identifier, whether the object 104 is experiencing an event within the sub-area and takes action in response to an event determined to be occurring within the sub-area.If the device 102 includes a microphone and the processor is coupled (e.g., mechanically, electrically, logically) to the microphone to control (e.g., receive data from) the microphone, the processor can assign an identifier to a sub-area via instructing (e.g., speaking) a user to output a sound corresponding to an identifier, such that the microphone captures sound as acoustic input and the processor determines whether an object is experiencing an event within the sub-area based on the set of acoustic inputs, the microphone captures sound as acoustic input and sends the acoustic input to the processor, the processor determines whether an object is experiencing an event within the sub-area based on the set of data and the identifier, and takes action in response to the event determined to be occurring within the sub-area. The processor identifies sub-areas within the area 100 based on the set of data, infers an area type (e.g., kitchen, bathroom) for the sub-area, classifies the sub-area based on the area type, and assigns an identifier (e.g., kitchen, bathroom). The processor assigns the identifier (e.g., kitchen, bathroom) to the sub-area based on the area type such that the processor determines whether the object 104 is experiencing an event within the sub-area based on the set of data and the area type, and takes action in response to the event determined to be occurring within the sub-area based on the identifier. A user can assign an identifier to the area 100 (e.g., via a user interface unit, a microphone unit, or a mobile app communicating with a communication unit of the device 102) such that the processor determines whether the object 104 is experiencing an event within the area 100 based on the set of data and the identifier, and takes action in response to the event determined to be occurring within the area 100.A user may assign an identifier to an area 100 by operating a communications unit (e.g., a mobile phone, a tablet computer, a wearable computer) in communication (e.g., wireless, waveguide, wired) with the device 102, such that the processor determines, based on the set of data and the identifier, whether an object 104 is experiencing an event within the area 100 and takes action in response to the event determined to be occurring within the area 100. If the device includes a microphone and the processor is coupled (e.g., mechanically, electrically, logically) to the microphone, the processor may also transmit audio signals (e.g., the microphone captures sound as acoustic input) that the user outputs (e.g., via a user interface unit, a speaker unit) to the processor, such that the processor determines, based on the set of data and the identifier, whether an object 104 is experiencing an event within the area 100 and takes action in response to the event determined to be occurring within the area 100.

[0099] As described herein, the device 102 is positioned within the area 100 for the radar to track objects 104 residing or located within the area 100. The radar can be placed anywhere within the area 100, but a preferred placement is at a corner or a predetermined distance from a corner. This offers various technical advantages. For example, radar antennas can be inexpensively constructed from flat circuit boards, but some antennas positioned on flat circuit boards cannot physically achieve a 180-degree field of view (although some can). In practice, this field of view 130 does not significantly exceed approximately 120 degrees horizontally without a 3D antenna structure. Therefore, positioning the radar at a corner maximizes or ensures that the most accurate portion of the radar's field of view 130 is aimed directly at the area of interest. Furthermore, radar operates by transmitting light over radio waves and measuring how the signal reflects off objects in space. If the radar were shone directly onto a wall, the wall would act as a prominent reflector, bouncing most of the radar's light back toward the radar and blinding it. Because many homes have some, many, most, or all of their walls built perpendicular to one another, placing the radar in a corner would place the center of the radar's field of view 130 at approximately a 45-degree angle relative to most walls, significantly reducing the amount of radar energy reflected by the walls. Furthermore, the processor can be programmed to recommend where to position the radar. For example, after the radar has operated within the area 100 for a predetermined amount of time, the processor can determine where the object 104 spends most of its time within the area 100. Based on the geofenced area in which the radar should operate and where people spend most of their time, the processor can recommend a different location that reduces the distance to the object 104 or reduces the number of obstacles between the radar and the object 104.Thus, if the area 100 has a corner, the processor can instruct the user (e.g., via a mobile app communicating with the communication unit of the device 102, via a user interface unit, a speaker unit) to position the device 102 at the corner or within a preset distance (e.g., within about 3, about 2, or about 1 feet or meter, or collinear with the corner, or so that the window 214 or its central partition is at about 45 degrees to the corner) of the area 100 having the object 104. The processor may be programmed to generate recommendations based on the data regarding where to position or reposition the device 102 or radar. The recommendations can optimize the minimum average spacing between the object 104 and the radar in the area 100, or the minimum number of occlusions or obstacles between the object 104 and the radar in the area 100. The recommendations can be output to the user (e.g., via a mobile app communicating with the communication unit of the device 102, via a user interface unit, a speaker unit).

[0100] As described herein, the processor is programmed to handle occlusions within the field of view 130. These occlusions introduce shadows (e.g., coverage gaps), which are areas where radar tracking is reduced or absent. However, even in these situations, the processor is able to recover as much information as possible. Regarding shadow measurement and representation, during geofencing, the reflectivity of the object 104 is recorded as the object 104 moves through the area 100, generating a 3D average reflectivity voxel map, with each voxel recording either the average reflectivity of the object 104 at that point or no detection. One way to approach this is to directly use the 3D average reflectivity model as a shadow representation. Voxels without direct measurements can take the reflectivity value of voxels close to the radar. Another approach is to plot polar coordinates that approximate the shadows cast by the occlusion. Such a frustrum is defined by five planes. Four planes can be plotted, each containing a point representing the radar center. One plane is orthogonal to the vector emanating from the center of the radar. Each plane can be calculated to align with an edge of the 3D average reflectance model, where the edge corresponds to a voxel transition from high reflectance to low reflectance. Edge detection can be performed using a 3D Sobel edge detector (or other suitable technique). Additionally, the processor can track occlusions. Using guidance from observations of the virtual skeleton's movement (e.g., running, walking, crawling), the velocity of the blob or skeleton's center of mass, the acceleration of the blob or skeleton's center of mass, and the room geometry, the processor can determine the likelihood that the object 104 has fallen behind an occlusion, as shown in FIG. 14. The movement of the virtual skeleton can be recorded by recording the xyz positions of skeleton keypoints over time, and the virtual blob is a primitive form of radar direction detection that detects significant reflectivity not due to static objects in the 3D voxel map of radar data, as shown in FIG. 14. For example, the radar observes the virtual skeleton keypoints, velocity, and acceleration of a person entering and exiting a shadow within X time (eg, about 15 seconds, about 30 seconds), as shown in FIG.This may not require an alert, but the processor can back-calculate the location of the person in the shadow. However, if the person does not move out of the shadow within Y time (e.g., about 20 seconds, about 40 seconds), the processor can choose to suppress the alert if the processor has previously observed this behavior. Otherwise, the processor can confirm the person's safety with an elevator or escalation alert. For example, the processor can request the speaker unit to play a sound asking if the person is safe, request the communication unit to contact the person via phone, or call / notify / message emergency contacts, or contact emergency services. The processor can handle occlusion based on the expression of a human motion model to infer activity patterns for prediction and classification. Three-axis spatial information, along with time-series timestamps, velocity, acceleration, rotation angle, and relative distance, represents the temporal and spatial displacement of moving body parts and is used to effectively handle occlusion or shadow and alert events. Furthermore, as described herein, using multiple devices 102, although overlapping is possible, can minimize these shadows by having non-overlapping fields of view 130 since their transmissions do not interfere.

[0101] As described herein, the device 102 may have a radar housed in a single housing, or the device 102 may have radar distributed across multiple housings, or multiple devices 102 may interoperate or operate in concert. For example, the device 102 may include a transmitter or transmitter antenna in a first housing and a receiver or receiver antenna in a second housing, which may be spaced apart from the first housing. These forms of distributed radar operation allow for one or more transmit antennas and one or more receive antennas to be located at different locations within the area 100, for example, by having such antennas in separate housings. This diversity of locations increases the field of view covered, improves the accuracy with which objects 104 are detected within the area 100 due to the different viewpoints of each transmitter / receiver pair, increases spatial information (e.g., more accurate measurements for different sides of a person / object), and helps cover sensor shadows. These antennas can be installed in a variety of ways. For example, there may be one set of transmit antennas in one housing and another set of receive antennas in another housing. For example, a first set of transmit antennas and a first set of receive antennas may be located in a first housing, and a second set of transmit antennas or a second set of receive antennas may be located in a second housing. For example, device 102 may include any combination of receive and transmit antennas grouped in a housing, as long as at least some transmissions of at least one transmit antenna can be received by at least one receive antenna. In some cases, various considerations may be necessary. For example, operating multiple physically separated transmit and receive antennas within area 100 may pose technical challenges. For example, if these antennas are not in a common housing and sensing occurs at the speed of light, coordinating how the antennas and their associated processors must communicate is significantly more complex than if the transmit and receive antennas were wired to a common processor. Furthermore, interference may sometimes need to be considered.For example, the transmit antenna of each radar may need to operate so as not to interfere with other radars. This can be achieved in several ways. For example, each transmit antenna can operate at a different frequency or bandwidth, or each transmit antenna can transmit a different digital pattern, such as transmitting a different Golay complementary pair. For example, each transmit antenna can transmit at a different time, so there are various ways to achieve time synchronization. For example, the first and second processors of the first and second devices 102 and 102 can synchronize their clocks (e.g., by designating one of the devices 102 to act as a master clock via a network time protocol over their communication units and communicating time content to each slave clock using a second communication channel, such as Z-Wave). In this way, each antenna can transmit during a predetermined window of time, each window of time being separated by a small period corresponding to a known synchronization error. For example, if each transmit antenna is paired with a receive antenna and shares a processor, the processor can preset the digital pattern for each radar, causing each radar to transmit in turn. Each radar transmits its pattern as it receives the pattern of the radar preceding it in sequence. Transmission begins with the first radar in the sequence transmitting its pattern without waiting to receive the pattern. Once all antennas have transmitted, the pattern loops. For example, if each transmit antenna is paired with a receive antenna and the processor is shared, a randomized backoff scheme can be used (e.g., as in code division multiple access (CDMA) or Wi-Fi transmission).

[0102] As described herein, radars can be installed in various locations within the area 100 for various use cases. The radars may be located by field strength. To comply with regulatory field strength limits, several transmitting antennas can measure the power of the surrounding electric fields and notify the user (e.g., via a user interface unit, a speaker unit, or a mobile app running on a mobile phone or tablet that communicates with the device 102's communication unit) when the field strength allows for locating the device 102. In one possible location scheme, a housing with a transmitting antenna is placed first and begins transmitting. Then, a receiving antenna or additional housings with transmitting antennas are placed. For these additional units, if the receiving antenna is paired with at least one transmitting antenna in the same housing, the receiving antenna can search for a field strength where the measured field strength and the expected field strength of the paired transmitting antenna do not exceed regulatory limits. For additional units, if the receiving antenna is not paired with at least one transmitting antenna in the same housing, the receiving antenna can measure the field strength and notify the user if the field is too weak to be detected. The radars can be located by a floor plan, where a floor plan of the area 100 can be provided to the devices 102, and the radars can be positioned to maximize coverage as guided by a user interface unit or speaker unit, or a mobile app running on a cell phone or tablet that communicates with the communication unit of the devices 102. The radars can be located by guidelines (e.g., specific or general), whereby the antennas can be positioned using guidelines such as approximate separation distance, height from the ground, distance from the ceiling, and forward-facing direction facing the inside of the house. The radars may also be located by a calculated floor plan / 3D map, and after or during this location process, each device 102 can calculate a floor plan based on a set of data from the radar: approximating the locations of floors, ceilings, and walls corresponding to large radar returns.The user can be presented with placement or recommended suggestions for improving placement based on the calculated floor plan (e.g., guided by a user interface unit or speaker unit, or a mobile app running on a mobile phone or tablet in communication with the device 102's communication unit). The radar can be positioned according to the observed user's location within the area 100 after the radar is initially positioned and after the radar observes several general locations of users within the area 100. The radar can suggest a new location that better covers the user's frequent locations. Thus, the radar can include a transmitter and a receiver, and a processor is coupled (e.g., mechanically, electrically, logically) to the transmitter and receiver for controlling them. The device 102 can include a first housing and a second housing, the first housing hosting the transmitter and the second housing hosting the receiver, and the first housing and the second housing being spaced apart from each other (e.g., within about 5, 4, 3, 2, or 1 feet or meters with an air gap between them). The first and second housings can be spaced apart from one another before the processor takes action in response to an event determined to be occurring within the area 100 based on the field strength of transmitters within the area 100, a received floor plan of the area 100, guidelines generic or specific to the area 100, a wizard generic or specific to the area 100, a calculated floor plan generic or specific to the area 100, a map generic or specific to the area 100, or an observed location of an object 104 within the area 100. The first and second housings may or may not be opposite each other and may include opposite corners. Further, there may be a first device 102 (a first housing hosting a first processor and a first radar having a first transmitter and a first receiver) and a second device 102 (a second housing hosting a second processor and a second radar having a second transmitter and a second receiver), where the first radar does not interfere with the second radar.The first and second housings may be spaced apart from each other or may be located at opposite corners, with the first transmitter configured to transmit a first signal receivable by the first and second receivers, and the second transmitter configured to transmit a second signal receivable by the first and second receivers. The first and second housings may be spaced apart from each other based on the field strength of the first or second transmitter within the defined area, a received floor plan of the defined area, a guideline general or specific to the defined area, a wizard general or specific to the defined area, a calculated floor plan general or specific to the defined area, a map general or specific to the defined area, or an observed location of an object within the defined area. The first and second processors may or may not be in communication with each other. The first and second radars may or may not overlap in the field of view 130.

[0103] As described herein, there may be situations in which the device 102 requires synchronization of scan patterns. For physically separated antennas to effectively operate as a radar system, each transmit and receive antenna pair needs to know with some precision how the other antenna is steering its transmit or receive beam. Coordination may be necessary to build a consistent model of the area 100. For example, if the processor's goal is to build a 3D voxel map of the area 100 as disclosed herein, the receive antenna may systematically scan the area 100, but if the transmit antenna is not transmitting in the same direction as the receive antenna is scanning, the receive antenna may record nothing or may fail to detect the object's reflectivity because there was no transmit antenna transmitting the digital pattern scattered from that object. Thus, in some use cases, there may be one transmit antenna in one housing and one receive antenna in another housing. For each transmit antenna steer (each beam direction), the receive antenna performs a complete scan of the area 100 (all possible beam steers). Coordination of when the transmitting antenna should switch beams and when the receiving antenna should scan the area 100 can be communicated via a communication unit (e.g., a Wi-Fi unit, a Bluetooth unit, a Zigbee unit, a Z-wave unit, or a cellular unit). In another use case, a pair of transmitting and receiving antennas in one housing faces another pair of transmitting and receiving antennas in another housing. The transmitting and receiving antennas used by the radars can also be used as a communication channel between the two radars. For example, one transmitting antenna steers a beam in one direction and transmits a pattern. After receiving a certain number of patterns, the receiving antenna repeats transmitting a different pattern. When the original antenna receives this alternative pattern, it stops transmitting the first pattern and begins receiving the alternative pattern. Physically separated radars need to know each other's beam steering patterns.This information can be communicated using a pre-configured radar antenna or another communication strategy via a communication unit (e.g., a Wi-Fi unit, Bluetooth unit, Zigbee unit, Z-wave unit, or cellular unit). If one radar fails to receive multiple transmissions from the other radar within a pre-configured time, it can restart the pattern using a different communication strategy, skip sections of the pattern where it cannot receive transmissions from the other radar, or abandon cooperation and operate as an independent radar unit. It is also possible to unify the data by further aggregating and processing the data collected from the separate radars. For example, if two radars enable a processor to detect a virtual skeleton, additional calculations can be used to determine whether the virtual skeletons represent the same physical person, and the measured coordinates of the virtual skeleton can be averaged across the measurements from the two separate radars, as shown in FIG. 14. The unification of data can occur in (a) a single housing where all radars transmit their measurement data to the same housing (via a transmitting antenna or via another communication method via a communication unit such as a Wi-Fi unit, Bluetooth unit, etc.); (b) a cloud computing service where all devices 102 transmit their measurement data to a cloud computing service; or (c) a distributed manner where each device 102 transmits its data to a subset of other devices 102 for further processing.

[0104] As described herein, the device 102 can be initially calibrated or configured, or recalibrated or reconfigured. One way this can happen is by the device 102 receiving various information about the attributes of the subject 104 (e.g., bone length, heart rate, respiration rate). As part of participant onboarding, the subject 104 is marked at a fixed distance from the radar with predictable movement to measure and mark bone length and center, usual walking pattern, usual walking speed, or other attributes so that a processor can create a persona or identifiable profile of the subject 104. For example, the subject 104 may be guided (e.g., by a user interface unit or speaker unit, or a mobile app running on a cell phone or tablet communicating with the device 102's communication unit) to sit still for a few seconds in front of the radar unit to measure the subject's 104's resting heart rate and respiration rate as part of the subject's 104 profile. Heart rate is extracted by processing the radar signal in the time domain and applying principal component analysis and other machine learning techniques to the time series data to reveal phase variations due to heartbeat. Respiration rate is measured using high-order harmonic peak selection and other machine learning methods. The recorded respiration rate or heart rate can be used as a baseline signature of the object 104 for subsequent learning. These vitals are captured for subsequent continuous tracking based on the identification of the object 104. Additionally, the processor can enable voice recognition setup of the object 104, and the processor can be programmed to obtain a voice profile of the object 104 to vocally identify, distinguish, supplement, or augment the object 104 from other objects 104 for radar tracking. The processor may also be programmed for identifying and tracking residential occupants within the area 100. For example, the processor may identify participants from objects 104 tracked within the area 100 and enable identification profiling of the objects 104 by utilizing the profile of the objects 104 when tracking vitals (e.g., resting heart rate, walking pattern, resting respiratory rate, depression index).For example, the processor may identify a new object 104 (e.g., a new resident residing or located in the area 100) by obtaining some details of this new resident for future reference or to track one or more participants with assistance from the participants. In this manner, the processor may be programmed to access a set of attributes of the object 104 and create a profile (or persona) of the object 104 based on the set of attributes, such that before taking action in response to an event determined to be occurring in the area 100, the processor determines whether the object 104 is experiencing an event in the area 100 based on the set of data and the profile. The profile may be a baseline from which the processor determines whether the object 104 is experiencing an event in the area 100 based on the set of data. If the device 102 includes a microphone and the processor is coupled (e.g., mechanically, electrically, logically) to the microphone, the processor may send an output (e.g., guided by a user interface unit or speaker unit, or a mobile app running on a mobile phone or tablet that communicates with a communication unit of the device 102) to the object 104 (e.g., the microphone captures acoustic input based on audio sounds, and the processor forms an audio profile of the object 104 and determines whether the object 104 is experiencing an event within the area 100 based on the set of data and the audio profile).

[0105] As described herein, the processor may request the radar to switch frequencies within the Ku, K, or Ka bands (e.g., to increase resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation), or the processor may request the radar to switch between at least two of the Ku, K, or Ka bands (e.g., The processor may also request the radar to switch modalities between Doppler and time-of-flight modes (e.g., to increase resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation). These requests can occur in a variety of ways.

[0106] For example, a single housing may contain multiple radars operating at different frequencies. One radar is Ku-band for detecting the location and attitude of objects 104 within the area 100, and the other is Ka-band for detecting vital signs of objects within the area 100. Based on a set of data from the radar operating at Ku-band, a processor can determine when to turn on or activate the Ka-band radar based on the location of a target (human or animal) within a known range where a heart rate can be reliably detected. In another approach, the processor can request the radar to shift its carrier frequency between Ku-band, K-band, or Ka-band depending on the desired operating mode (e.g., Ku-band for location tracking, Ka-band for vital signs, or both = K-band) to conserve power, control heat dissipation, and control other operating or tracking (e.g., accuracy, precision, resolution) parameters. Using a set of high-frequency components (e.g., beamformers, up / down converters, oscillators) can reduce costs at the expense of operating outside or near the boundaries of these components' frequency capabilities. For radars that can shift frequencies within the Ku, K, or Ka bands, the high-frequency components have a desired operating range, and operating outside this frequency range may require more power to transmit at the same field strength. The choice of which frequencies to use within the Ku, K, or Ka bands, or how to switch frequencies within the Ku, K, or Ka bands, can also be based on achieving a desired power usage profile. Additionally, switching frequencies within the Ku, K, or Ka bands may be necessary if high-frequency components become too hot because they are being pushed to operate at frequencies near or outside their design range.Thus, the radar may be configured to operate in the Ku band or the Ka band, and the processor may be programmed to activate the radar operating in the Ku band or the Ka band within the area 100 to track a residence or object 104 located within the area 100, generate a set of data based on tracking the residence or object 104 located within the area 100 in the Ku band or the Ka band, and transmit the set of data to the processor so that the processor determines whether the object 104 is experiencing an event within the area 100 based on the set of data and takes action in response to the event determined to be occurring within the area 100. The processor may switch the radar between the K band and the Ku band or the Ka band based on parameters that meet or do not meet thresholds (e.g., increase resolution of the object 104 or the area 100 or its contents, conserve power, control heat dissipation). Note that the radar is configured to operate in parallel in the K band and the Ku band or the Ka band without interfering with each other. When there are at least two devices 102 interacting or cooperating within the area 100, one of these devices 102 may have a processor switch bands between Ku-band, K-band, or Ka-band based on a parameter that meets or does not meet a threshold (e.g., increasing resolution of the object 104 or area 100 or its contents, conserving power, controlling heat dissipation), or a parameter that meets or does not meet a threshold (e.g., increasing resolution of the object 104 or area 100 or its contents, conserving power, controlling heat dissipation). These two radars may or may not be spaced apart from each other within the area 100 and may be located at corners of the area 100. As a result, these two devices 102 may or may not identify the same or different events and may or may not take the same or different actions. It should be noted that whether tracking in Ku-band, K-band, or Ka-band, the two radars may or may not operate in parallel, or may or may not interfere with each other.

[0107] For example, the radar may switch modalities between Doppler mode (or another radar modality) and time-of-flight mode (or another radar modality) when requested by the processor based on the processor's determination that various criteria, signatures, or thresholds have or have not been met (e.g., to enhance resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation), as disclosed herein. Note that such switching may or may not operate in series or parallel, may or may not interfere with each other, and may or may not be accompanied by frequency switching or band switching, regardless of whether the radar is operating in Ku band, K band, Ka band, or other band, as disclosed herein. For example, a radar may have a first radar unit operating in Doppler mode and a second radar unit operating in time-of-flight mode, with the processor requesting that the first radar unit operate in Doppler mode and then switch to the second radar unit to operate in time-of-flight mode, or vice versa, based on the processor determining whether various criteria, signatures, or thresholds have been met or not (e.g., to enhance resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation, as disclosed herein, although parallel or serial radar mode operation is also possible). Note that the first and second radar units may be hosted by a common housing (e.g., internally, externally), as disclosed herein, or each may have its own housing that may be spaced apart (e.g., within about 5, 4, 3, 2, 1 feet, or meters) from one another.Similarly, for example, the radar may be capable of parallel or serial radar mode operation, but may also be operated in Doppler or time-of-flight mode, where the processor requests the radar to operate in Doppler mode and then switch to time-of-flight mode, or vice versa, based on the processor determining whether various criteria, signatures, or thresholds have been met or not (e.g., to enhance resolution of the object 104 or area 100 or its contents, or to manage power or heat dissipation), as disclosed herein.

[0108] Various embodiments of the present disclosure may be implemented in a data processing system suitable for storing and / or executing program code, including at least one processor coupled directly or indirectly to memory elements via a system bus. The memory elements may include, for example, local memory used during the actual execution of the program code, bulk storage, and cache memory that provides temporary storage of at least some of the program code to reduce the number of times the code must be retrieved from bulk storage during execution.

[0109] I / O devices (including but not limited to keyboards, displays, pointing devices, DASDs, tapes, CDs, DVDs, thumb drives, and other memory media) may be coupled to the system either directly or through intervening I / O controllers. Network adapters may also be coupled to the system to enable the data processing system to become coupled to other data processing systems or remote printers or storage devices through intervening private or public networks. Modems, cable modems, and Ethernet cards are just a few of the types of network adapters available.

[0110] The present disclosure may be embodied in a system, a method, and / or a computer program product. The computer program product may include a computer-readable storage medium (or medium) having computer-readable program instructions thereon for causing a processor to execute aspects of the present disclosure. The computer-readable storage medium may be a tangible device capable of holding and storing instructions for use by an instruction execution device. The computer-readable storage medium may be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination thereof. A non-exhaustive list of more specific examples of computer-readable storage media includes portable computer diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disk read-only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanically encoded devices such as punch cards or raised structures in grooves with instructions recorded thereon, and any suitable combination thereof.

[0111] The computer-readable program instructions described herein can be downloaded to each computing / processing device from a computer-readable storage medium or to an external computer or external storage device over a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can be composed of copper transmission cables, fiber optic transmission cables, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage on a computer-readable storage medium in each computing / processing device.

[0112] Computer-readable program instructions for carrying out operations of the present disclosure can be either source code or object code written in any combination of one or more programming languages, including assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microphone code, firmware instructions, state-setting data, or object-oriented programming languages such as Smalltalk, C++, and conventional procedural programming languages such as the "C" programming language or similar programming languages. A code segment or machine-executable instructions can represent a procedure, a function, a subprogram, a program, a routine, a subroutine, a module, a software package, a class, or any combination of instructions, data structures, or program statements. A code segment can be coupled to another code segment or a hardware circuit by passing information, data, arguments, parameters, or memory contents. Information, arguments, parameters, data, etc. can be passed, forwarded, or transmitted via any suitable means, including memory sharing, message passing, token passing, network transmission, etc. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection may be to an external computer (e.g., via the Internet using an Internet Service Provider).In various embodiments, electronic circuitry including, for example, a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA) can utilize state information of the computer-readable program instructions to execute the computer-readable program instructions and personalize the electronic circuitry to carry out aspects of the present disclosure.

[0113] Aspects of the present disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions. The various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.

[0114] The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowcharts or block diagrams may represent a module, segment, or portion of instructions, which may consist of one or more executable instructions for implementing specified logical functions. In some alternative embodiments, the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks shown in succession may in fact be executed substantially concurrently, or the blocks may be executed in the reverse order, depending on the functionality involved. It should also be noted that each block of the block diagrams and / or flowchart diagrams, and combinations of blocks in the block diagrams and / or flowchart diagrams, may be implemented by a special-purpose hardware-based system that performs the specified functions or acts, or a combination of special-purpose hardware and computer instructions.

[0115] The use of words such as "then," "next," and the like is not intended to limit the order of the steps; these words are simply used to guide the reader through the description of the method. Although a process flow diagram may describe operations as sequential processes, many operations may be performed in parallel or concurrently. Additionally, the order of operations may be rearranged. A process corresponds to a method, a function, a procedure, a subroutine, a subprogram, and the like. When a process corresponds to a function, its termination corresponds to the return of the function to the calling function or the main function.

[0116] Although various embodiments have been illustrated and described in detail herein, those skilled in the art will recognize that various modifications, additions, substitutions, etc., may be made without departing from the present disclosure, and therefore, such modifications, additions, substitutions, etc. are considered to be within the scope of the present disclosure. [Explanation of symbols]

[0117] 102 devices Stand 106 108 sofa 110 Tables 112 chairs 114 Oven 116 Refrigerator 120 Toilet 122 TUB 124 Sink 126 outlets 128 Wall 132 Wall 136 Windows 134, 138 doors

Claims

1. It is a device, A device comprising a housing that hosts a processor, an artificial intelligence (AI) accelerator, a time-of-flight (TOF) radar, and a communication interface, which can be placed within a defined area having an object inside, the housing being capable of (a) the TOF radar detecting the object based on its respiration rate, tracking the object within the defined area, generating a set of data based on tracking the object within the defined area, and transmitting the set of data to the processor, (b) the processor forming a three-dimensional skeletal model based on the set of data, (i) enabling the AI ​​accelerator to determine, based on the set of data and the three-dimensional skeletal model, whether the object is experiencing an event within the defined area, and (ii) enabling the communication interface to take action based on the AI ​​accelerator's determination that the object is experiencing the event within the defined area based on the set of data and the three-dimensional skeletal model.

2. The device according to claim 1, wherein the AI ​​accelerator capacity-determines, based on the set of data, whether the object has experienced the event within the defined area.

3. The TOF radar is the device according to claim 1, which operates in the K-band.

4. The device according to claim 1, wherein the object is a first object, the set of data is a first set of data, the TOF radar tracks a second object within the defined area, generates a second set of data based on the tracking of the second object within the defined area, and the processor can transmit the second set of data to the processor so that the AI ​​accelerator can distinguish the first object from the second object and determine, based on the set of data, whether the object is experiencing the event within the defined area.

5. The device according to claim 4, wherein the second object is a pet.

6. The device according to claim 1, wherein the housing can host a set of microphones and receive a set of acoustic inputs generated from the objects in the defined area in order to enable the processor to confirm or verify the set of data.

7. The device according to claim 1, wherein the set of data is a first set of data, (a) the TOF radar can track the object outside the defined area, generate a second set of data based on tracking the object outside the defined area, and transmit the second set of data to the processor, and (b) the processor is programmed to discard, remove, delete or ignore the second set of data.

8. The device according to claim 1, wherein the event relates to the object that remains stationary within the defined area for a predetermined period of time.

9. The device according to claim 1, wherein the event is related to the object being absent from the defined area for a predetermined period of time.

10. The device according to claim 1, wherein the event relates to the object not being tracked within the defined area for a predetermined period of time while the object is within the defined area.

11. The device according to claim 1, wherein the defined area has corners, and the processor is programmed to generate an output relating to where to position or reposition the housing before the action.

12. The device according to claim 1, wherein the event is a diagnostic estimation or diagnostic prediction.

13. The device according to claim 1, wherein the TOF radar has a field of view of approximately 120 degrees horizontally and approximately 90 degrees vertically.

14. The device according to claim 1, wherein the TOF radar includes a set of phased arrays, each having a set of patch antennas.

15. The device according to claim 1, wherein the processor is programmed to access a set of attributes of the object before the action and to create a profile of the object based on the set of attributes before the action, and the AI ​​accelerator determines, based on the set of data and the profile, whether the object is experiencing the event within the defined area.

16. The device according to claim 1, wherein the TOF radar is actively cooled.

17. The device according to claim 1, wherein the TOF radar is passively cooled.

18. The defined area is located inside a building, and the device is as described in claim 1.

19. The device according to claim 1, wherein the housing hosts at least one of an accelerometer, a gyroscope, a compass, a light sensor, a temperature sensor, a humidity sensor, or a particulate sensor.

20. It is a method, Users The method includes arranging a housing that hosts a processor, an artificial intelligence (AI) accelerator, a time-of-flight (TOF) radar, and a communication interface within a defined area having an object inside, and (a) the TOF radar can detect the object based on its respiration rate, track the object within the defined area, generate a set of data based on tracking the object within the defined area, and transmit the set of data to the processor. (b) The processor forms a three-dimensional skeletal model based on the set of data; (i) the AI ​​accelerator determines, based on the set of data and the three-dimensional skeletal model, whether the object is experiencing an event within the defined area; and (ii) the communication interface takes action based on the AI ​​accelerator's determination that the object is experiencing the event within the defined area.

Images