Method and system for identifying exceptions of people behavior
A method and system for identifying exceptions, of a person/child behavior through scheduled behavior, using a personal communication device having at least one sensor attached to person/child body which provide location data and motion data. The method includes sampling measurements data for each sensor of the personal communication device, identifying atomic pattern of sampled measurements, the atomic pattern representing basic behavior of the person/child, identify location in which person has spent time using a clustering algorithm, create activity segmentation characterized by location, schedule, caregiver, or content derived from sensors measurements and atomic patterns and analyzing characteristics changes or combination thereof of activities in sequenced/complex activities in comparison to baseline, to identify exception which indicate of at least one of the following: changes at activity level, changes at emotional arousal level, unknown locations or unexpected locations based on schedule.
1. A method of remotely monitoring the emotional state of a cognitively disabled patient, using a personal communications computing device associated with at least a location data sensor, a motion data sensor, and a microphone attached to the patient's body or garments, said method comprising the steps of:
sampling measurements from the location and motion data sensors;
identifying atomic patterns of the sampled motion measurements and identifying a time and a geographic location in which the atomic patterns are detected by analyzing real time basic behavior pattern and comparing the identified time and geographic location of the atomic pattern to known pattern templates of classified action;
receiving and storing in a computer readable storage medium information indicative of typical behavior from a guardian associating at least one specific atomic pattern or at least one specific time interval with at least one specific emotion of the cognitively disabled patient, the received information identifying stress from the atomic pattern or indicating the specific interval represents stressful behavior;
using a clustering algorithm generated by computer-implemented supervised learning to classify activity segments in real time characterized by the location, the time, and the stored information received from the guardian, the classification resulting in at least a first activity segment characterized by a first location and a second activity segment characterized by a second location, the first and second locations being different from one another;
comparing at least one of characteristics changes and irregular activities in the classified activity segments with a baseline to identify an exception indicating an emotional state;
responsively to identifying the exception, activating the microphone and sampling input from the microphone to identify one or more keywords related to the cognitively disabled patient provided by the guardian; and
responsively to the identification of the one or more keywords, reporting an emotional arousal alert.
2. The method of claim 1 further comprising the step of sending an alert indicating a severity of the exception.
3. The method of claim 1 wherein the activity segmentation further includes classifying transitions between locations.
4. A system for remotely monitoring the emotional state of a cognitively disabled patient, the system comprising a personal communications computing device associated with at least a location data sensor, a motion data sensor, and a microphone attached to the patient's body or garments, the system configured to perform the steps of:
sampling measurements from the location and motion data sensors;
identifying atomic patterns of the sampled motion measurements and identifying a time and a geographic location in which the atomic patterns are detected by analyzing a real-time basic behavior pattern and comparing the identified time and geographic location of the atomic pattern to known pattern templates of classified actions;
receiving and storing in a computer readable storage medium information indicative of typical behavior from a guardian associating at least one specific atomic pattern or at least one specific time interval with at least one specific emotion of the cognitively disabled patient, the received information identifying stress from the atomic pattern or indicating the specific interval represents stressful behavior;
using a clustering algorithm generated by computer-implemented supervised learning to classify activity segments in real time characterized by the location, the time, and the stored information received from the guardian, the classification resulting in at least a first activity segment characterized by a first location and a second activity segment characterized by a second location, the first and second locations being different from one another;
comparing at least one of characteristics changes and irregular activities in the classified activity segments with a baseline to identify an exception indicating an emotional state;
responsively to identifying the exception, activating the microphone and sampling input from the microphone to identify one or more keywords related to the cognitively disabled patient provided by the guardian; and
responsively to the identification of the one or more keywords, reporting an emotional arousal alert.
5. The system of claim 4 , wherein the personal communications computing device is further configured for sending an alert indicating a severity of the exceptions.