IP Library › Granted Patent US 12,367,976
Granted Patent B1
US 12,367,976 · App. 18/589,262 · Granted Jul 22, 2025

Detecting falls with multiple wearable motion sensors

Inventors: Denis Filatov (Mexico City, MX); Hitten Zaveri (New Haven, CT)
Assignee: Alva Health Inc.
G16H40/67G06F1/163
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,367,976
App. No.
18/589,262
Granted
Jul 22, 2025
Kind
B1
Abstract

In an example system, multi-sensor motion data that reflects the motion of a user over a period of time is received as recorded by a plurality of motion sensors on wearable devices, specific changes are determined in the data by, firstly, quantifying it using a two-dimensional data transform; secondly, extracting an anomalous area; and thirdly, calculating a number of the anomaly's properties, and the results are input into a machine learning model to detect that the user fell. The machine learning model processes the anomaly's properties and evaluates the current state of the user's activity by classifying the properties against a state space previously calculated by analyzing historical activities of daily living. Depending on a two-dimensional transform implemented, the machine learning model detects when the user falls, as well as potentially allows predicting that the user will suffer a fall in advance of the actual event, aiming at solving the stroke prediction problem.

Claims (50)

1. A method comprising:

receiving, at a plurality of wearable devices worn by a user, multi-sensor motion data based on the user's activities over a period of time, wherein each of the plurality of wearable devices includes a motion sensor, wherein the plurality of wearable devices includes a first wearable device worn on a first wrist of the user and a second wearable device worn on a second wrist of the user;

calculating one or more properties of the multi-sensor motion data, wherein the calculating of the one or more properties of the multi-sensor motion data comprises:

filtering time series of the multi-sensor motion data;

calculating a measure of coherence of the filtered multi-sensor motion data;

determining, using the obtained measure of coherence, an area, wherein the coherence's values exceed a threshold; and

calculating one or more properties of the area;

determining that the user fell using a machine learning model executed on a processor of a cloud service and based on the one or more properties of the multi-sensor motion data, wherein the machine learning model was previously trained on historical multi-sensor motion data; and

transmitting the determination to a server, a caretaker of the user, and an emergency system.

2. The method of claim 1 , wherein the machine learning model was previously trained either on a person-specific basis, using historical multi-sensor motion data associated with the user's past activities, or on a universal basis, using historical multi-sensor motion data associated with past activities of a cohort of users, wherein the historical multi-sensor motion data were recorded using the motion sensors of the plurality of wearable devices.

3. The method of claim 1 , wherein the multi-sensor motion data recorded by the plurality of motion sensors includes one- to three-component vectors measuring acceleration of parts of the user's body over the period of time at a rate of at least 20 Hz.

4. The method of claim 1 , wherein the filtering of the time series of the multi-sensor motion data comprises calculating a two-dimensional time-frequency data transform followed by averaging a result in frequency and differentiating in time.

5. The method of claim 1 , wherein the measure of coherence is indicative of joint time-frequency features of the time series of the multi-sensor motion data after the filtering.

6. The method of claim 1 , wherein the one or more properties are indicative of joint time-frequency features of the area.

7. The method of claim 1 , wherein using the machine learning model executed on the processor further comprises evaluating a state of the user's current activity.

8. The method of claim 7 , wherein the evaluating of the state of the user's current activity is computed by classifying the one or more properties of an area against a previously determined optimal state space based on an analysis of the historical multi-sensor motion data.

9. A computer program product embodied in a non-transitory computer readable medium and comprising computer instructions for:

receiving, at a plurality of wearable devices worn by a user, multi-sensor motion data based on the user's activities over a period of time, wherein each of the plurality of wearable devices includes a motion sensor, wherein the plurality of wearable devices includes a first wearable device worn on a first wrist of the user and a second wearable device worn on a second wrist of the user;

calculating one or more properties of the multi-sensor motion data, wherein the calculating of the one or more properties of the multi-sensor motion data comprises:

filtering time series of the multi-sensor motion data;

calculating a measure of coherence of the filtered multi-sensor motion data;

determining, using the obtained measure of coherence, an area, wherein the coherence's values exceed a threshold; and

calculating one or more properties of the area;

determining that the user fell using a machine learning model executed on a processor of a cloud service and based on the one or more properties of the multi-sensor motion data, wherein the machine learning model was previously trained on historical multi-sensor motion data; and

transmitting the determination to a server, a caretaker of the user, and an emergency system.

10. The computer program product of claim 9 , wherein the machine learning model was previously trained either on a person-specific basis, using historical multi-sensor motion data associated with the user's past activities, or on a universal basis, using historical multi-sensor motion data associated with past activities of a cohort of users, wherein the historical multi-sensor motion data were recorded using the motion sensors of the plurality of wearable devices.

11. The computer program product of claim 9 , wherein the multi-sensor motion data recorded by the plurality of motion sensors includes one- to three-component vectors measuring acceleration of parts of the user's body over the period of time at a rate of at least 20 Hz.

12. The computer program product of claim 9 , wherein the filtering of the time series of the multi-sensor motion data comprises calculating a two-dimensional time-frequency data transform followed by averaging a result in frequency and differentiating in time.

13. The computer program product of claim 9 , wherein the measure of coherence is indicative of joint time-frequency features of the time series of the multi-sensor motion data after the filtering.

14. The computer program product of claim 9 , wherein the one or more properties are indicative of joint time-frequency features of the area.

15. The computer program product of claim 9 , wherein using the machine learning model executed on the processor further comprises evaluating a state of the user's current activity.

16. The computer program product of claim 15 , wherein the evaluating of the state of the user's current activity is computed by classifying the one or more properties of an area against a previously determined optimal state space based on an analysis of the historical multi-sensor motion data.

17. A system comprising:

a processor configured to:

receive, at a plurality of wearable devices worn by a user, multi-sensor motion data based on the user's activities over a period of time, wherein each of the plurality of wearable devices includes a motion sensor, wherein the plurality of wearable devices includes a first wearable device worn on a first wrist of the user and a second wearable device worn on a second wrist of the user;

calculate one or more properties of the multi-sensor motion data, wherein the calculating of the one or more properties of the multi-sensor motion data comprises to:

filter time series of the multi-sensor motion data;

calculate a measure of coherence of the filtered multi-sensor motion data;

determine, using the obtained measure of coherence, an area, wherein the coherence's values exceed a threshold; and

calculate one or more properties of the area;

determine that the user fell using a machine learning model executed on a processor of a cloud service and based on the one or more properties of the multi-sensor motion data, wherein the machine learning model was previously trained on historical multi-sensor motion data; and

transmit the determination to a server, a caretaker of the user, and an emergency system; and

a memory coupled to the processor and configured to provide the processor with instructions.

18. The system of claim 17 , wherein the machine learning model was previously trained either on a person-specific basis, using historical multi-sensor motion data associated with the user's past activities, or on a universal basis, using historical multi-sensor motion data associated with past activities of a cohort of users, wherein the historical multi-sensor motion data were recorded using the motion sensors of the plurality of wearable devices.

19. The system of claim 17 , wherein the multi-sensor motion data recorded by the plurality of motion sensors includes one- to three-component vectors measuring acceleration of parts of the user's body over the period of time at a rate of at least 20 Hz.

20. The system of claim 17 , wherein the filtering of the time series of the multi-sensor motion data comprises calculating a two-dimensional time-frequency data transform followed by averaging a result in frequency and differentiating in time.

21. The system of claim 15 , wherein the measure of coherence is indicative of joint time-frequency features of the time series of the multi-sensor motion data after the filtering.

22. The system of claim 17 , wherein the one or more properties are indicative of joint time-frequency features of the area.

23. The system of claim 17 , wherein using the machine learning model executed on the processor further comprises to evaluate a state of the user's current activity.

24. The system of claim 23 , wherein the evaluating of the state of the user's current activity is computed by classifying the one or more properties of an area against a previously determined optimal state space based on an analysis of the historical multi-sensor motion data.

Continuity (1)
Continuation 18202213 · May 25, 2023
References Cited (34)
US 11020059B2 · Sheth · 2021 [cited by applicant]
US 11282361B2 · Sharma · 2022 [cited by applicant]
US 20150006446A1 · Chiba · 2015 [cited by applicant]
US 20190347922A1 · Burton · 2019 [cited by applicant]
US 20200205746A1 · Burwinkel · 2020 [cited by applicant]
US 20210174663A1 · Widgren · 2021 [cited by applicant]
US 20230000396A1 · Coffey · 2023 [cited by applicant]
US 20230125403A1 · Chia-Lin · 2023 [cited by applicant]
Adhikari et al., Activity Recognition for Indoor Fall Detection Using Convolutional Neural Network, 2017 Fifteenth IAPR International Conference on Machine Vision Applications (MVA), IEEE, May 8, 2017, pp. 81-84. [cited by applicant]
Bat-Erdene et al., Automatic Acute Stroke Symptom Detection and Emergency Medical Systems Alerting by Mobile Health Technologies: A Review, Journal of Stroke and Cerebrovascular Diseases, vol. 30, No. 7, Jul. 2021, pp. … [cited by applicant]
Blanchard, Efficacy of a smartwatch application in detecting induced falls, Submitted to JMIR mHealth and uHealth, unpublished, non-peer-reviewed preprint, May 2, 2021, 20 pages, https://preprints.jmir.org/preprint/3012… [cited by applicant]
Botonis et al., Wearable airbag technology and machine learned models to mitigate falls after stroke, Journal of NeuroEngineering and Rehabilitation, 19(1):60, Jun. 17, 2022, pp. 1-14, https://doi.org/10.1186/s12984-022… [cited by applicant]
Chelli et al., A Machine Learning Approach for Fall Detection and Daily Living Activity Recognition, IEEE Access, vol. 7, Mar. 21, 2019, pp. 38670-38687. [cited by applicant]
Filatov, On spatial synchronisation as a manifestation of irregular energy cascades in continuous media under the transition to criticality, Nonlinear Dynamics, 109(4), Jun. 16, 2022, pp. 2573-2594, https://doi.org/10.1… [cited by applicant]
Hakim et al., Smartphone Based Data Mining for Fall Detection: Analysis and Design, ScienceDirect, Procedia Computer Science 105, Jan. 1, 2017, pp. 46-51. [cited by applicant]
Harris, Mstroke: Methods of Fall Detection and Data Storage, A Thesis Submitted to the Faculty of the University of Tennessee at Chattanooga in Partial Fulfillment of the Requirements of the Degree of Master of Science:… [cited by applicant]
Ihlen et al., The Discriminant Value of Phase-Dependent Local Dynamic Stability of Daily Life Walking in Older Adult Community-Dwelling Fallers and Nonfallers, BioMed Research International, Article ID 402596, Oct. 2015… [cited by applicant]
Jahanjoo et al., Accurate Fall Detection Using 3-Axis Accelerometer Sensor and MLF Algorithm, 3rd International Conference on Pattern Recognition and Image Analysis (IPRIA 2017), Apr. 19-20, 2017, pp. 90-95. [cited by applicant]
Kao et al., GA-SVM applied to the fall detection system, Proceedings of the 2017 IEEE International Conference on Applied System Innovation (ICASI), May 13, 2017, pp. 436-439. [cited by applicant]
Kostopoulo et al., F2D: A location aware fall detection system tested with real data from daily life of elderly people, ScienceDirect, Procedia Computer Science 98, Jan. 1, 2016, pp. 212-219. [cited by applicant]
Liang et al., Fall Detection using Lifting Wavelet Transform and Support Vector Machine, Proceedings of the Federated Conference on Computer Science and Information Systems, ISSN 2300-5963 ACSIS, vol. 11, Sep. 3, 2017, … [cited by applicant]
Liu et al., Impact of Sampling Rate on Wearable-based Fall Detection Systems Based on Machine Learning Models, IEEE Sensors Journal 18, No. 23, Sep. 30, 2018, pp. 9882-9890. [cited by applicant]
Lv et al., Information collection system for fall detection of stroke patients under cascade algorithm in the context of multi-modal information fusion and e-health, Expert Systems 40, No. 4, May 2023, pp. 1-11, https:/… [cited by applicant]
Özdemir et al., Detecting Falls with Wearable Sensors Using Machine Learning Techniques, Sensors 2014, No. 14, ISSN 1424-8220, Jun. 18, 2014, pp. 10691-10708. [cited by applicant]
Putra et al., An Event-Triggered Machine Learning Approach for Accelerometer-Based Fall Detection, Sensors 2018, No. 1, Dec. 22, 2017, pp. 1-18. [cited by applicant]
Ralhan, A Study on Machine Learning Algorithms for Fall Detection and Movement Classification, PhD diss., Master's thesis, Department of Electrical and Computer Engineering, University of Saskatchewan, Dec. 2009, 67 pag… [cited by applicant]
Ramachandran et al., A Survey on Recent Advances in Wearable Fall Detection Systems, BioMed Research International, vol. 2020, Article ID 2167160, Jan. 13, 2020, pp. 1-17, https://doi.org/10.1155/2020/2167160. [cited by applicant]
Ramachandran et al., Machine Learning-based Fall Detection in Geriatric Healthcare Systems, 2018 IEEE International Conference on Advanced Networks and Telecommunications Systems (ANTS), Dec. 16, 2018, pp. 1-6. [cited by applicant]
Taylor-Piliae et al., Objective fall risk detection in stroke survivors using wearable sensor technology: a feasibility study, Topics in Stroke Rehabilitation, ISSN: 1074-9357 (Print) 1945-5119 (Online), May 10, 2016, 8… [cited by applicant]
Tran et al., A multi-modal multi-view dataset for human fall analysis and preliminary investigation on modality, 2018 24th International Conference on Pattern Recognition (ICPR), Aug. 20-24, 2018, pp. 1947-1952. [cited by applicant]
Wang et al., An Improved Fall Detection Approach for Elderly People Based on Feature Weight and Bayesian Classification, Proceedings of 2016 IEEE International Conference on Mechatronics and Automation, Aug. 7-10, 2016,… [cited by applicant]
Weiss et al., Does the Evaluation of Gait Quality During Daily Life Provide Insight Into Fall Risk? A Novel Approach Using 3-Day Accelerometer Recordings, Neurorehabilitation and Neural Repair 27, No. 8, Oct. 2013, pp. … [cited by applicant]
Yang et al., A Wearable Real-Time Fall Detector Based on Naive Bayes Classifier, in CCECE 2010 IEEE, May 2, 2010, pp. 1-4. [cited by applicant]
Yu et al., Hidden Markov Model-Based Fall Detection with Motion Sensor Orientation Calibration: A Case for Real-Life Home Monitoring, IEEE Journal of Biomedical and Health Informatics 22, No. 6, Dec. 11, 2017, pp. 1847-… [cited by applicant]