Granted Patent B2

US 12,525,110 · App. 18/524,352 · Granted Jan 13, 2026

Inventors: Dongeek Shin (San Jose, CA); Shwetak Patel (Seattle, WA); Rizwan Chaudhry (Millbrae, CA); Chetan Bhole (Mountain View, CA); Vaibhav Darbari (New York, NY); Todd Whitehurst (Belmont, CA); Anupam Pathak (San Carlos, CA)

Assignee: Google LLC

G08B21/043G01S13/32G01S13/58G08B21/0469

Abstract

Claims (77)

1 . A device, comprising:

radar circuitry comprising an antenna array for emitting radar waves and receiving reflected radar waves; and

a processing system, comprising one or more processors, that is in communication with the radar circuitry, the processing system being configured to:

while operating in an idle state in which fall monitoring is not performed, determine that exactly one person is present in an environment of the device, wherein fall monitoring is disabled when more than one person is present;

in response to determining that exactly one person is present, transition to a fall monitoring state;

while in the fall monitoring state with exactly one person present:

receive waveform data from the radar circuitry;

perform center-of-mass tracking on a person present within the waveform data;

determine that a fall by the person has occurred based on performing the center-of-mass tracking; and

output, via a speaker, speech announcing the fall;

receive an auditory response via a microphone; and

perform an action in response to the auditory response, wherein the action is selected from the group consisting of:

transmitting a communication via a wireless network to an emergency contact; and

outputting a spoken message indicating that no further action is to be taken.

2 . The device of claim 1 , wherein:

the processing system is further configured to determine that the person is present for monitoring; and

the radar circuitry emitting radar waves is performed in response to determining that the person is present for monitoring.

3 . The device of claim 1 , wherein the radar circuitry emits frequency-modulated continuation wave (FMCW) radio waves.

4 . The device of claim 1 , wherein the processing system is further configured to:

classify the person based on the center-of-mass tracking.

5 . The device of claim 4 , wherein classifying the person based on the center-of-mass tracking is performed using a pre-trained machine learning model.

6 . The device of claim 5 , wherein the pre-trained machine learning model as installed on the device is static and was created based on a training set of data indicative of situations in which persons have fallen and in which persons have not fallen.

7 . The device of claim 1 , wherein the processing system being configured to perform center-of-mass tracking comprises the processing system being configured to perform center-of-mass tracking over a rolling historic time period window.

8 . The device of claim 1 , wherein the radar circuitry comprises a plurality of antennas.

9 . The device of claim 8 , wherein the processing system being configured to receive the waveform data comprises the processing system being configured to receive a plurality of individual sets of raw waveform data from the plurality of antennas of the radar circuitry.

10 . The device of claim 1 , wherein the processing system is further configured to:

perform feature extraction based on the center-of-mass tracking, wherein the feature extraction comprises extracting a plurality of features selected from the group consisting of:

initial azimuthal position;

final azimuthal position;

azimuthal position change;

azimuthal slope;

initial elevational position;

final elevational position;

elevational position change;

elevational slope;

initial range position;

final range position;

range position change;

range slope;

initial RCS (radar cross section) position;

final RCS position;

RCS position change;

RCS slope; and

velocity.

11 . A method for performing fall detection, the method comprising:

while operating in an idle state in which fall monitoring is not performed, determining, by a device, that exactly one person is present in an environment of the device, wherein fall monitoring is disabled when more than one person is present;

in response to determining that exactly one person is present, transitioning, by the device, to a fall monitoring state;

while in the fall monitoring state with exactly one person present:

emitting, by the device, radar waves, wherein the device comprises:

radar circuitry comprising an antenna array for transmitting radar waves and receiving reflected radar waves;

based on reflected radar waves, creating raw waveform data;

filtering, from the raw waveform data, waveform data indicative of static objects to obtain motion-indicative waveform data;

performing, by the device, center-of-mass tracking on a person present within the motion-indicative waveform data;

determining, by the device, that a fall by the person has occurred based on performing the center-of-mass tracking; and

outputting, by the device via a speaker of the device, a first spoken message announcing that the fall was determined to have occurred.

12 . The method of claim 11 , further comprising:

receiving an auditory response via a microphone and process the auditory response; and

performing an action in response to the auditory response, wherein the action is selected from the group consisting of:

transmitting a communication via a wireless network to an emergency contact; and

outputting a spoken message indicating that no further action is to be taken.

13 . The method of claim 11 , wherein the radar waves emitted are frequency-modulated continuation wave (FMCW) radio waves.

14 . The method of claim 11 , further comprising:

classifying the person based on the center-of-mass tracking.

15 . The method of claim 14 , wherein classifying the person based on the center-of-mass tracking is performed using a pre-trained machine learning model.

16 . A non-transitory processor-readable medium, comprising processor-readable instructions configured to cause one or more processors of a device to:

while operating in an idle state in which fall monitoring is not performed, determine that exactly one person is present in an environment of the device, wherein fall monitoring is disabled when more than one person is present;

in response to determining that exactly one person is present, transition to a fall monitoring state;

while in the fall monitoring state with exactly one person present:

receive raw waveform data from radar circuitry;

filter, from the raw waveform data, waveform data indicative of static objects to obtain motion-indicative waveform data;

perform center-of-mass tracking on a moving object present within the motion-indicative waveform data;

determine that a fall by a person has occurred based on performing the center-of-mass tracking; and

output, via a speaker, speech announcing that the fall has been detected;

receive an auditory response via a microphone and process the auditory response; and

perform an action in response to the auditory response, wherein the action is selected from the group consisting of:

causing a communication to be transmitted via a wireless network to an emergency contact; and

outputting, via the speaker, a spoken message indicating that no further action is to be taken.

Continuity (2)

References Cited (102)