Automatic Recognition, Learning, Monitoring, and Management of Human Physical Activities

• US 20150224362A1
• Filed: 02/12/2015
• Published: 08/13/2015
• Est. Priority Date: 02/12/2014
• Status: Active Grant

• Current Assignee: Khaylo Inc.
• Original Assignee: Khaylo Inc.
• Inventors: Rishi, Alok, Rishi, Arjun, Moran, Alex, Vanungare, Niranjan
• Granted Patent: US 10,022,071 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: A61B 5/0205 Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
• CPC Class Codes: A61B 5/021 Measuring pressure in heart or blood vessels A61B5/0205 takes precedence
• CPC Class Codes: A61B 5/02438 with portable devices, e.g. worn by the patient
• CPC Class Codes: A61B 5/1118 Determining activity level
• CPC Class Codes: A61B 5/1123 Discriminating type of movement, e.g. walking or running A61B5/1116, A61B5/112 take precedence
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: G16H 20/30 relating to physical therapies or activities, e.g. physiotherapy, acupressure or exercising
• CPC Class Codes: G16H 20/40 relating to mechanical, radiation or invasive therapies, e.g. surgery, laser therapy, dialysis or acupuncture

Abstract

In one embodiment, a method includes one or more processors that collect motion and physiological sensor data of a user from one or more sensors worn by the user, the sensor data comprising one or more sensor data vectors, generate an activity signature based on the sensor data, determine whether a signature match exists between the activity signature and a known activity signature associated with an activity type from a set of known activity signatures, and if the signature match exists, recognize a known activity type, otherwise if the signature match does not exist, generate an unknown activity type based on the one or more sensor data vectors.

Claims

1. A method comprising:

by one or more processors;

collect motion and physiological sensor data of a user from one or more sensors worn by the user, the sensor data comprising one or more sensor data vectors;

generate an activity signature based on the sensor data;

determine whether a signature match exists between the activity signature and a known activity signature associated with an activity type from a set of known activity signatures;

if the signature match exists, recognize a known activity type;

if the signature match does not exist, generate an unknown activity type based on the one or more sensor data vectors.

2. The method of claim 1, wherein determine whether a signature match exists comprises performing a vector quantization process on the activity signature.

3. The method of claim 1, further comprising classify the unknown activity type for future matching.

4. The method of claim 1, wherein generate an unknown activity type comprises machine learning techniques.

5. The method of claim 1, further comprising provide user feedback to the user based on the known activity type or the unknown activity type.

6. The method of claim 1, further comprising provide user feedback to the user based on the known activity type or the unknown activity type in a context-aware manner.

7. The method of claim 6, wherein the context-aware manner comprises a social network, sports network, fitness network, or health network.

8. The method of claim 1, further comprising:

compare the known or unknown activity type with a contemporaneously scheduled activity in an activity calendar of the user;

mark the contemporaneously scheduled activity as completed.

9. The method of claim 8, further comprising provide user feedback to a second user based on the completed contemporaneously scheduled activity.

10. A system comprising:

one or more processors operable to;

collect motion and physiological sensor data of a user from one or more sensors worn by the user, the sensor data comprising one or more sensor data vectors;

generate an activity signature based on the sensor data;

determine whether a signature match exists between the activity signature and a known activity signature associated with an activity type from a set of known activity signatures;

if the signature match exists, recognize a known activity type;

if the signature match does not exist, generate an unknown activity type based on the one or more sensor data vectors.

11. The system of claim 10, further comprising classify the unknown activity type for future matching.

12. The system of claim 10, further comprising provide user feedback to the user based on the known activity type or the unknown activity type.

13. The system of claim 10, further comprising provide user feedback to the user based on the known activity type or the unknown activity type in a context-aware manner.

14. The system of claim 10, further comprising

compare the known or unknown activity type with a contemporaneously scheduled activity in an activity calendar of the user;

mark the contemporaneously scheduled activity as completed.

15. The system of claim 14, further comprising provide user feedback to a second user based on the completed contemporaneously scheduled activity.

16. One or more computer-readable non-transitory storage media embodying software that is operable when executed to:

collect motion and physiological sensor data of a user from one or more sensors worn by the user, the sensor data comprising one or more sensor data vectors;

generate an activity signature based on the sensor data;

determine whether a signature match exists between the activity signature and a known activity signature associated with an activity type from a set of known activity signatures;

if the signature match exists, recognize a known activity type;

if the signature match does not exist, generate an unknown activity type based on the one or more sensor data vectors.

17. The media of claim 16, further comprising classify the unknown activity type for future matching.

18. The media of claim 16, further comprising provide user feedback to the user based on the known activity type or the unknown activity type.

19. The media of claim 16, further comprising provide user feedback to the user based on the known activity type or the unknown activity type in a context-aware manner.

20. The media of claim 16, further comprising

compare the known or unknown activity type with a contemporaneously scheduled activity in an activity calendar of the user;

mark the contemporaneously scheduled activity as completed.