Activation of Ancillary Sensor Systems Based on Triggers from a Wearable Gesture Sensing Device

Abstract

An event detection system includes sensors to detect movement and other physical inputs related to a user, which the event detection system can process to identify gestures of the user, and possibly also determine, using historical data, machine learning, rule sets, or other techniques for processing data to derive an inferred event related to the user sensed by the sensors. An inferred event might be an eating event, a smoking event, a personal hygiene event, a medication related event, or some other event the user is inferred to be engaging in. When an event is inferred to have started, to be ongoing, and/or to have concluded, the event detection system can take actions related to that event, such as obtaining other information to be stored in memory in association with the data representing the event, interacting with the user to provide information or reminders or to prompt for user input, sending a message to a remote computer system, sending a message to another person, such as a friend, health care provider, first responder, or other action(s).

What is claimed is: 
     
         1 . An event detection system that includes sensors to detect movement and/or other physical inputs related to a user, which the event detection system can process to identify gestures of the user, a method of further processing comprising:
 determining a type of an inferred event, wherein the inferred event is an actual event occurring with or about the user or is, based on sensor inputs, deemed by the event detection system to be an event;   determining an external trigger time, based on one or more of data related to the inferred event, the type of the inferred event, and/or a timing of the inferred event, wherein the external trigger time is a time determined relative to an anchor time of the inferred event; and   according to the external trigger time, initiating a computer-based action in response to the inferred event.   
     
     
         2 . The event detection system of  claim 1 , wherein the sensors include sensors external to a main event detection unit. 
     
     
         3 . The event detection system of  claim 1 , wherein the computer-based action in response to the inferred event comprises sending a trigger signal to an ancillary system that is operable independent of the event detection system and independent of the trigger signal. 
     
     
         4 . The event detection system of  claim 1 , wherein the event detection system is configured to determine, using historical data, machine learning, rule sets, and/or other techniques for processing data, an inferred event related to the user sensed by the sensors. 
     
     
         5 . The method of  claim 1 , wherein the type of the inferred event is at least one of a food intake event, a drinking event, a smoking event, a personal hygiene event, and/or a medication related event. 
     
     
         6 . The method of  claim 1 , wherein the external trigger time is determined from when the inferred event is inferred to have started, when the inferred event is inferred to be ongoing, and/or when the inferred event is inferred to have concluded. 
     
     
         7 . The method of  claim 1 , wherein the computer-based action in response to the inferred event is one or more of (1) obtaining other information to be stored in memory in association with data representing the inferred event, (2) interacting with the user to provide information, coaching advice or a reminder, (3) interacting with the user to prompt for user input, (4) sending a message to a remote computer system, (5) sending one or more inputs to a medication dispensing system, and/or (6) sending a message to another person. 
     
     
         8 . The method of  claim 7 , wherein the computer-based action comprises inputs to the medication dispensing system and wherein the medication dispensing system is an automated, or partially automated, insulin delivery system. 
     
     
         9 . The method of  claim 7 , wherein the medication dispensing system comprises an insulin dosage calculator. 
     
     
         10 . The method of  claim 9 , wherein the insulin dosage calculator computes information from an object information retrieval subsystem about what is being consumed. 
     
     
         11 . The method of  claim 1 , further comprising updating one or more of an inventory database, a medication log, an inventory, and/or a production line database in response to the inferred event. 
     
     
         12 . The method of  claim 1 , wherein the computer-based action comprises an action to retrieve information about at least one object the user is interacting with. 
     
     
         13 . The method of  claim 12 , wherein the retrieved information is information retrieved over a wireless link. 
     
     
         14 . The method of  claim 13 , wherein the wireless link uses one of (1) near-field-communication technology, (2) Bluetooth technology, (3) Bluetooth Low Energy technology, (4) a derivative of Bluetooth technology that is at least partially consistent with Bluetooth technology, or (5) a derivative of Bluetooth Low Energy technology that is at least partially consistent with Bluetooth Low Energy technology.
 A method of managing a food log based in part on sensor data from sensors on a device worn by a user for logging the user's food intake is also provided, that might include identifying a current eating event; further identifying one or more characteristics of the current event, from the current eating event and the characteristics of the current event, and from historical data of past eating events, using a training system to identify additional characteristics of a current eating event; and periodically uploading data representing characteristics of eating events to a food log server.   
     
     
         15 . A method of managing a food log based in part on sensor data from sensors on a device worn by a user for logging the user's food intake, the method comprising:
 identifying a current eating event;   identifying one or more characteristics of the current eating event;   from the current eating event and the characteristics of the current event, and from historical data of past eating events, using a training system to identify additional characteristics of a current eating event; and   periodically uploading data representing characteristics of eating events to a food log server.   
     
     
         16 . A method of managing a food log based in part on sensor data from sensors on a device worn by a user for logging the user's food intake, the method comprising:
 identifying a current eating event;   identifying one or more characteristics of the current eating event;   periodically uploading data representing characteristics of eating events to a food log server;   comparing uploaded characteristics of food intake events to second set of eating characteristics that have been obtained through a different mechanism; and   as a result of comparing, take one or more of the following actions (1) merge characteristics from both datasets correspond to the same food intake event into a single eating event data record of a food logging system, and (2) adding entries to the food log for the events absent from the food log but represented by identified characteristics of the current eating event.   
     
     
         17 . The method of  claim 15 , further comprising:
 using historical data of past eating events to identify additional characteristics of a current eating event; and   uploading additional characteristics to a food log server.   
     
     
         18 . The method of  claim 15 , wherein periodically uploading data comprises periodically uploading data at conclusions of detected eating events. 
     
     
         19 . The method of  claim 15 , wherein periodically uploading data comprises uploading data at times that correspond to anchor times of detected eating events. 
     
     
         20 . The method of  claim 15 , wherein the data representing characteristics of eating events is derived from inferences derived from sensed user movements during eating events. 
     
     
         21 . The method of  claim 15 , wherein the data representing characteristics of eating events is partially inferred from sensed user movements during eating events and meal signatures are added to the food log so as to allow the user to edit the food log while being reminded of the meal signature. 
     
     
         22 . The method of  claim 15 , wherein recording the inferred event in the food log comprising recording one or more of information inferred by the event detection system and information retrieved from one or more objects the user interacted with as part of the inferred event, wherein user interaction is detected using sensors on a device worn by a user and sensor signals from those sensors or from devices designed to read electronic tags on the user or the food. 
     
     
         23 . The method of  claim 15 , wherein the food log comprises information about eating or drinking events including one or more of a time of an event, a duration of the event, a location of the event, metrics related to pace of consumption during the event, metrics related to quantities consumed during the event, eating methods, and/or utensils used. 
     
     
         24 . The method of  claim 15 , wherein the food log comprises information about items being eaten or drunk, wherein such information is obtained from electronically readable tags attached to the items, the method further comprising sending a tag reader trigger signal to a tag reader to read the electronically readable tags when an inferred event is detected. 
     
     
         25 . The method of  claim 24 , further comprising prompting the user to reposition the items and the tag reader to a relative position in which the tag reader can read the electronically readable tags attached to the items. 
     
     
         26 . A method of sending signals or messages in response to an event timed relative to the event, the method comprising:
 identifying behavior indicators;   from the behavior indicators and timestamps of the behavior indicators, and from historical data of past behavior events and behavior indicators, using a training system to predict likely future events including a time of such likely future events; and   reading a rule set;   when the rule set indicates that a particular signal should be sent when a particular event is anticipated and when the particular event is indeed anticipated with at least a predefined confidence level, outputting the particular signal to an ancillary system; and   when the rule set indicates that a particular message should be sent when the particular event is anticipated and when the particular event is indeed anticipated with at least the predefined confidence level, sending the particular message.   
     
     
         27 . The method of  claim 26 , wherein identifying behavior indicators comprises:
 detecting at least one gesture of a user wearing a wearable device having a plurality of sensors to detect movement and other physical inputs related to a user;   receiving sensor inputs from the plurality of sensors;   reading data from external data sources; and   from the sensors inputs and the data, identifying the behavior indicators.   
     
     
         28 . The method of  claim 26 , wherein the particular event is an eating event, wherein the particular signal is a signal to trigger an insulin microdosing system and the particular message is a message to the user of a wearable device indicating that the start of the eating event has been detected 
     
     
         29 . The method of  claim 26 , wherein the particular event is such that it could not be predicted from the sensor inputs alone or the data from external sources alone. 
     
     
         30 . The method of  claim 26 , wherein the particular event is an eating event and behavior indicators correlate with triggers, wherein triggers include one or more of a sleep pattern, a stress level, an activity level, the user's location, people surrounding the person, vital signs, a hydration level, a fatigue level, or a heart rate. 
     
     
         31 . The method of  claim 26 , wherein the signal or message is a function of a level of user engagement. 
     
     
         32 . The method of  claim 31 , further comprising assessing the level of user engagement at a given time or over a time window, combined with user responses to certain messages.