Fall identification system

1. A method comprising:

receiving, by a computing device, sensor data from a motion sensing device associated with a user;

storing the sensor data in a buffer memory;

identifying a user context, the user context including an identified pre-fall activity based on the sensor data;

after identifying the user context, identifying a potential fall based on motion sensing data read from the buffer memory;

detecting a presence of body micro movements of the user after the potential fall;

determining, based on a continued presence of the detected body micro movements after the potential fall, that the motion sensing device is currently worn by the user;

in response to determining that the motion sensing device is currently worn by the user, assigning a probability of the potential fall being an actual fall based on data from an additional sensor; and

updating the probability of the potential fall being the actual fall based on the user context, the user context including the identified pre-fall activity, wherein the updating the probability of the potential fall being the actual fall further comprises:

utilizing a machine learning system to determine a probability of a false positive identification for the potential fall based on the identified pre-fall activity, the machine learning system trained on cumulative training data from a plurality of users' systems, the cumulative training data comprising pre-fall activities of the plurality of users, and the machine learning system being trained to output the probability of a false positive identification for a particular pre-fall activity, and

updating the determined probability to an updated probability based on the probability of false positive identifications and the identified pre-fall activity of the user.

2. The method of claim 1 , further comprising:

in response to the updated probability exceeding a threshold, sending a fall alert to another device.

3. The method of claim 1 , further comprising:

in response to determining that the motion sensing device is a body-worn device not currently worn by the user, classifying the potential fall as a dropped device.

4. The method of claim 1 , wherein the data from the additional sensor comprises one of: barometric data, gyroscopic data, or sound data.

5. A method of determining whether a user has fallen comprising:

detecting a potential fall using a motion sensing device worn by the user;

identifying a pre-fall activity of the user prior to the potential fall;

detecting, by the motion sensing device, body micro movements of the user;

determining, based on a continued presence of the detected body micro movements after the potential fall, that the motion sensing device that detected the body micro movements is a body-worn device currently worn by the user; and

in response to determining that the motion sensing device is the body-worn device currently worn by the user, evaluating the potential fall to determine whether the potential fall is an actual fall based on user context, the user context including the identified pre-fall activity, the evaluating comprising:

utilizing a machine learning system to determine a probability of a false positive identification for the potential fall based on the identified pre-fall activity, the machine learning system trained on cumulative training data from a plurality of users' systems, the cumulative training data comprising pre-fall activities of the plurality of users, and the machine learning system being trained to output the probability of the false positive identification for a particular pre-fall activity, and

calculating a probability that the potential fall was an actual fall based on the probability of the false positive identifications and the identified activity of the user.

6. The method of claim 5 , wherein the evaluating utilizes data from an additional sensor.

7. The method of claim 5 , further comprising receiving input from a plurality of additional sensors and wherein the evaluating comprises sequentially evaluating the plurality of additional sensors until the probability of an actual fall meets a threshold.

8. The method of claim 7 , wherein the plurality of additional sensors comprise one or more of: gyroscope, barometer, microphone, blood sugar sensor, blood pressure sensor, and blood oxygenation sensor.

9. The method of claim 5 , wherein the identified activity comprises one of: sitting, standing, jogging, bicycling, and skateboarding.

10. A method of determining whether a user has fallen comprising:

detecting a potential fall based on data from a body-worn motion sensing device worn by the user;

detecting, by the motion sensing device, body micro movements of the user;

determining, based on a continued presence of the detected body micro movements after the potential fall, that the motion sensing device is currently worn by the user;

in response to determining that the motion sensing device is the body-worn device is currently worn by the user:

assigning a probability to a potential fall based on user context, the user context including an identified pre-fall activity prior to the potential fall,

updating the probability of the potential fall being an actual fall based on data from an additional sensor,

wherein the updating the probability of the potential fall being an actual fall further comprises:

utilizing a machine learning system to determine a probability of a false positive identification for the potential fall based on the identified pre-fall activity, the machine learning system trained on cumulative training data from a plurality of users' systems, the cumulative training data comprising pre-fall activities of the plurality of users, and the machine learning system being trained to output the probability of a false positive identification for a particular pre-fall activity, and

updating the determined probability to an updated probability based on the probability of false positive identifications and the identified pre-fall activity of the user.

11. The method of claim 10 , wherein the additional sensor is located in an additional device.

12. The method of claim 10 , wherein the additional sensor comprises a barometer.

13. The method of claim 10 , wherein the additional sensor comprises a microphone.

14. The method of claim 10 , wherein the additional sensor performs one of: blood sugar analysis, blood pressure analysis, and blood oxygenation analysis.

15. The method of claim 10 , wherein the identified pre-fall activity comprises one of: sitting, standing, jogging, bicycling, and skateboarding.

16. The method of claim 10 , wherein the probability is a percentage value between 0 and 99%.

17. The method of claim 10 , wherein updating the probability of the potential fall being an actual fall based on user context comprises:

determining a false positive indication probability for the identified pre-fall activity prior to the potential fall, and adjusting the probability based on the false positive indication probability.

18. The method of claim 10 , further comprising:

utilizing location and environmental context data to validate the potential fall.

19. The method of claim 10 , further comprising sequentially evaluating a plurality of additional sensors until the probability of an actual fall meets a threshold.