Using health monitor data to detect macro and micro habits with a behavioral model

• US 9,750,433 B2
• Filed: 12/16/2014
• Issued: 09/05/2017
• Est. Priority Date: 05/28/2013
• Status: Active Grant

• Current Assignee: LARK Technologies, Inc. (Abbvie Incorporated)
• Original Assignee: LARK Technologies, Inc. (Abbvie Incorporated)
• Inventors: Hu, Julia, Zira, Jeff
• Primary Examiner(s): Stout, Michael C
• Assistant Examiner(s): Kolderman, Nicholas E
• Application Number:
  US14/572,601
• Publication Number:
  US 20160007912A1
• Time in Patent Office: 994 Days
• CPC Class Codes: A61B 2560/0242 adapted to measure environmental factors, e.g. temperature, pollution
• CPC Class Codes: A61B 5/002 Monitoring the patient using a local or closed circuit, e.g. in a room or building A61B5/0017 takes precedence
• CPC Class Codes: A61B 5/1112 Global tracking of patients, e.g. by using GPS
• 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 5/6824 Arm or wrist
• CPC Class Codes: A61B 5/6898 Portable consumer electronic devices, e.g. music players, telephones, tablet computers
• CPC Class Codes: A61B 5/725 using specific filters therefor, e.g. Kalman or adaptive filters A61B5/316 take precedence
• CPC Class Codes: A61B 5/7282 Event detection, e.g. detecting unique waveforms indicative of a medical condition A61B5/4094, A61B5/4818, A61B5/349, A61B5/0823 take precedence
• CPC Class Codes: H04L 12/1859 adapted to provide push services, e.g. data channels
• CPC Class Codes: H04L 51/214 using selective forwarding
• CPC Class Codes: H04L 51/216 Handling conversation history, e.g. grouping of messages in sessions or threads

Abstract

A method for communicating activity-related notifications to a user includes: receiving a record of activity events of a particular activity type performed by the user over a period of time; selecting a first time-based filter from a set of time-based filters; identifying a cluster of activity events in the record of activity events filtered according to the first time-based filter; identifying an early bound and a late bound on start times of activity events of the particular activity type from the cluster; communicating a notification of a first type to the user at a first time within a threshold time of the early bound on a day fulfilling the first time-based filter; and communicating a notification of a second type to the user at a second time within a threshold time of the late bound on a day fulfilling the first time-based filter.

Claims

1. A method for communicating activity-related notifications to a user, the method comprising:

receiving, at a remote computer system from a wireless communication transceiver module of a mobile computing device, a record of activity events of a particular activity type performed by the user over a period of time, the record specifying a start time and a duration of each activity event in the record of activity events, wherein receiving the record of activity events comprises collecting location data sampled at a location sensor of the mobile computing device comprising the location sensor, a display, and the wireless communication transceiver module, wherein the location data comprises a location coincident performance of each activity event by the user;

selecting, at the remote computer system, a first time-based filter from a set of time-based filters for the record;

identifying, at the remote computer system, a cluster of activity events in the record of activity events filtered according to the first time-based filter, wherein identifying the cluster of activity events in the record of activity events comprises;

obtaining a set of computer-implemented rules defining activity event subgrouping as a function of activity event start time, activity event duration, and coincident location, wherein the set of computer-implemented rules are operable to improve the remote computing system in relation to behavioral pattern prediction, and

selecting a subgroup of activity events according to the set of computer-implemented rules and corresponding to;

composite start times and durations falling within a core composite distance threshold of a composite start time and a duration of at least one other activity event within the subgroup of activity events, and

locations, from the location data sampled at the location sensor, within a threshold distance of a location corresponding to at least one other activity event within the subgroup of activity events; and

predicting, at the remote computer system, representation of a behavioral pattern of the user in the subgroup based on identification of the cluster;

identifying, at the remote computer system, an early bound on start times of activity events of the particular activity type from the cluster of activity events;

identifying, at the remote computer system, a late bound on start times of activity events of the particular activity type from the cluster of activity events;

at a first time succeeding the period of time, communicating a notification of a first type to the user for presentation at the display of the mobile computing device, the first time within a threshold time of the early bound on a day fulfilling the first time-based filter; and

at a second time succeeding the period of time, communicating a notification of a second type to the user for presentation at the display of the mobile computing device, the second time within a threshold time of the late bound on a day fulfilling the first time-based filter.

2. The method of claim 1, wherein receiving the record of activity events comprises collecting a first set of action data of the user from a wearable device worn by the user during the period of time, collecting a second set of action data of the user from a mobile computing device associated with the user during the period of time, fusing the first set of action data and the second set of action data into discrete activity events performed by the user during the period of time, and filtering the discrete activity events by the particular activity type to assemble the record of activity events.

3. The method of claim 1, wherein receiving the record of activity events comprises retrieving, from a database, a record of walking events performed by the user over a sequence of days, the record specifying a start time and a duration of each walking event in the record of walking events;

wherein selecting the first time-based filter from the set of time-based filters comprises selecting the first time-based filter specifying a minimum duration of a walking event and a particular subset of weekdays; and

wherein identifying the cluster of activity events in the record of activity events comprises discarding walking events, in the record of walking events, of duration less than the minimum duration and occurring on days outside the particular subset of weekdays.

4. The method of claim 3, wherein selecting the first time-based filter comprises selecting the first time-based filter from the set of time-based filters comprising a filter specifying a single weekday, a filter specifying a combination of two weekdays, a filter specifying a combination of three weekdays, a filter specifying a combination of four weekdays, a filter specifying a combination of five weekdays, a filter specifying a single weekend day, and a filter specifying a combination of weekend days.

5. The method of claim 1, wherein identifying the cluster of activity events in the record of activity events comprises:

identifying activity events in the subgroup as core points;

selecting a second subgroup of activity events, in the record of activity events, corresponding to composite start times and durations exceeding the core composite distance threshold and falling within a density-reachable threshold distance of a composite start time and duration of an activity event within the subgroup of activity events;

identifying activity events in the second subgroup as density-reachable points;

calculating a centroid of core points and density-reachable points in the subgroup and the second subgroup;

calculating a maximum distance and a minimums distance between the centroid and the core points and between the centroid and the density-reachable points; and

predicting representation of a minimum of two behavioral patterns in the subgroup and the second subgroup in response to a difference between the maximum distance and the minimum distance exceeding a threshold.

6. The method of claim 1,

further comprising, from the cluster of activity events, identifying a particular geographic location associated with the activity events of the particular activity type;

further comprising setting a threshold distance from the particular geographic location for activity events of the particular activity type for the user;

wherein communicating the notification of the first type to the user comprises communicating the notification of the first type through a mobile computing device associated with the user based on presence of the computing device within the threshold distance of the particular geographic location on a day fulfilling the first time-based filter.

7. The method of claim 1, wherein identifying the cluster of activity events in the record of activity events comprises identifying a subgroup of activity events, in the record of activity events, corresponding to start times falling within a threshold time from a start time of at least one other activity event within the subgroup of activity events.

8. The method of claim 1, further comprising identifying a user behavioral pattern for the particular activity type based on the subgroup of activity events comprising a number of activity events exceeding a threshold number of events.

9. The method of claim 8, wherein identifying the user behavioral pattern of the particular activity type comprises characterizing the subgroup of activity events as one of a pre-habit behavior and a habit based on the number of activity events in the subgroup of activity events;

and wherein communicating the notification of the first type to the user comprises generating the notification of the first type based on a characterization of the subgroup of activity events as one of a pre-habit behavior and a habit.

10. The method of claim 1, wherein communicating the notification of the first type to the user comprises extrapolating, from the cluster, a quantitative datum corresponding to the particular activity type and presenting a form of the quantitative datum through a mobile computing device associated with the user.

11. The method of claim 10, wherein extrapolating the quantitative datum from the cluster comprises calculating a probability of initiation of an activity of the particular activity type, by the user, between the early bound and the late bound on a current date and predicting a duration of the activity of the particular activity type based on the cluster of activity events;

and wherein presenting the form of the quantitative datum comprises rendering the probability of initiation of the activity and the duration of the activity within in a notification on a display of the mobile computing device in response to instance of a time within a threshold time of the early bound on a day fulfilling the first time-based filter.

12. The method of claim 10, wherein communicating the notification of the second type to the user comprises, in response to an absence of a detected activity of the particular activity type between the early bound and the late bound on a day fulfilling the first time-based filter, presenting to the user, through the mobile computing device, a prompt to provide feedback for the absence of a detected activity of the particular activity type.

13. The method of claim 12, further comprising:

receiving feedback for the absence from the user through the mobile computing device;

selecting an alternative time-based filter from the set of time-based filters based on the feedback;

identifying a second cluster of activity events in the record of activity events filtered according to the alternative time-based filter;

identifying an alternative early bound on start times of activity events of the particular activity type from the second cluster of activity events;

identifying an alternative late bound on start times of activity events of the particular activity type from the alternative cluster of activity events;

at a third time succeeding the period of time, communicating a notification of the first type to the user, the third time within a threshold time of the alternative early bound on a day fulfilling the alternative time-based filter; and

at a fourth time succeeding the period of time, communicating a notification of the second type to the user, the fourth time within a threshold time of the alternative late bound on a day fulfilling the alternative time-based filter.

14. The method of claim 1, further comprising:

receiving a second record of activity events of the particular activity type performed by the user over a second period of time succeeding the period of time, the second record specifying a start time and a duration of each activity event in the second record of activity events;

selecting a second time-based filter from a set of time-based filters for the record and the second record;

identifying a second cluster of activity events in a combination of the record of activity events and the second record of activity events filtered according to the second time-based filter, the second cluster comprising at least one activity event of the cluster;

identifying an alternative early bound on start times of activity events of the particular activity type from the second cluster of activity events;

identifying an alternative late bound on start times of activity events of the particular activity type from the second cluster of activity events;

at a third time succeeding the second period of time, communicating a notification of the first type to the user, the third time within a threshold time of the alternative early bound on a day fulfilling the second time-based filter; and

at a fourth time succeeding the second period of time, communicating a notification of the second type to the user, the fourth time within a threshold time of the alternative late bound on a day fulfilling the second time-based filter.

15. The method of claim 1:

wherein receiving the record of activity events comprises receiving the record further specifying a second activity event preceding each activity event in the record of activity events by a corresponding lead time;

wherein identifying the cluster of activity events in the record of activity events comprises identifying a subgroup of activity events, in the record, preceded by second activities of a second activity type corresponding to lead times less than a maximum lead time specified in the first time-based filter, the second activity type distinct from the first activity type; and

wherein communicating the notification of the first type to the user comprises communicating the notification of the first type to the user in response to detecting an activity of the second type prior to the early bound on a day fulfilling the first time-based filter.

16. The method of claim 1,

wherein receiving the record of activity events comprises collecting motion data sampled at an inertial sensor of the mobile computing device, wherein the motion data describes user motion coincident the performance of each activity event by the user, wherein the location data and the motion data are operable to improve the remote computer system in relation to activity type prediction, and the method further comprising:

classifying, at the remote computer system, the particular activity type performed by the user based on the location data and the motion data, wherein predicting representation of the behavioral pattern comprises predicting representation of the behavioral pattern based on classification of the particular activity type and the identification of the cluster.

17. A method for communicating activity-related notifications to a user, the method comprising:

receiving, at a remote computer system from a wireless communication transceiver module of a mobile computing device, a record of activity events of a particular activity type performed by the user over a period of time, the record specifying a time of each activity event in the record of activity events, wherein receiving the record of activity events comprises collecting location data sampled at a location sensor of the mobile computing device comprising the location sensor, a display, and the wireless communication transceiver module, wherein the location data comprises a location coincident performance of each activity event by the user;

selecting, at the remote computer system, a first time-based filter from a set of time-based filters for the record;

identifying, at the remote computer system, a cluster of activity events in the record of activity events filtered according to the first time-based filter, wherein identifying the cluster of activity events in the record of activity events comprises;

obtaining a set of computer-implemented rules defining activity event subgrouping as a function of activity event start time, activity event duration, and coincident location, wherein the set of computer-implemented rules are operable to improve the remote computing system in relation to behavioral pattern prediction, and

selecting a subgroup of activity events according to the set of computer-implemented rules and corresponding to;

composite start times and durations falling within a core composite distance threshold of a composite start time and a duration of at least one other activity event within the subgroup of activity events, and

locations, from the location data sampled at the location sensor, within a threshold distance of a location corresponding to at least one other activity event within the subgroup of activity events; and

predicting, at the remote computer system, representation of a behavioral pattern of the user in the subgroup based on identification of the cluster;

identifying, at the remote computer system, an early bound on times of activity events of the particular activity type from the cluster of activity events;

identifying, at the remote computer system, a late bound on times of activity events of the particular activity type from the cluster of activity events;

extrapolating, from the cluster, a quantitative datum corresponding to the particular activity type and the first time-based filter;

at a first time succeeding the period of time, presenting a form of the quantitative datum through the display of the mobile computing device associated with the user, the first time within a threshold time of the early bound on a day fulfilling the first time-based filter; and

at a second time succeeding the period of time, in response to an absence of a detected activity of the particular activity type between the early bound and the late bound on a day fulfilling the first time-based filter, presenting to the user a prompt to provide feedback for the absence of a detected activity of the particular activity type, the second time within a threshold time of the late bound on a day fulfilling the first time-based filter.

18. The method of claim 17, further comprising:

receiving feedback for the absence from the user through the mobile computing device;

selecting an alternative time-based filter from the set of time-based filters based on the feedback;

identifying a second cluster of activity events in the record of activity events filtered according to the alternative time-based filter;

identifying an alternative early bound on times of activity events of the particular activity type from the second cluster of activity events; and

identifying an alternative late bound on times of activity events of the particular activity type from the alternative cluster of activity events.

19. The method of claim 17, wherein identifying the cluster of activity events in the record of activity events comprises identifying a minimum number of activity events, in the record of activity events, corresponding to start times falling within a range of start times specified in the first time-based filter for a particular weekday.

20. The method of claim 17, wherein receiving the record of activity events of the particular activity type comprises receiving the record of activity events of the particular activity type comprising one of walking, exercising, eating, driving, working, and sleeping, the record further specifying a duration of each activity event in the record of activity events.