Automatic in-home senior care system augmented with internet of things technologies
The in-home care of seniors is augmented using Internet of Things (IOT) technologies. In-home sensors monitor a senior and their caregiver. Physical conditions and psychological conditions may be monitored. In some implementations, a machine learning system has a classifier trained to detect a specified condition, such as depression. The system may perform various transformations of raw sensor data into a format indicative of a particular condition. In one implementation, a psychological or medical condition has symptoms in which each symptom has one or more measurable events. Mappings between symptoms, events, sensor data, and sensor transformation functions may be supported.
1. A system for providing in-home care for seniors, comprising:
a machine learning subsystem including a processor configured to:
generate a normalized feature vector from a set of sensor outputs from sensors disposed within a living area of a senior according to a sensor floorplan, including identifying a first relationship between symptoms of depression and events, wherein the events are measurable physical or mental features associated with at least one of the symptoms of depression, identifying a second relationship between the events and the set of sensor outputs based at least in part on the locations of each sensor and a sensor type of each sensor; and identifying a third relationship of sensor transformations required to transform sensor data into a format indicative of events, wherein the first relationship, the second relationship, and the third relationship is used to generate the normalized feature vector;
input the normalized feature vector to a logistic regression classifier of a machine learning model, wherein the logistic regression classifier is trained to determine thresholds for identifying depression based on a training data set of a set of seniors; and
determine a likelihood that the senior has depression.
2. The system of claim 1 , wherein the machine learning subsystem is further configured to generate an alert in response to determining a risk the senior is deprressed.
3. The system of claim 1 , further comprising a sensor installation subsystem to determine a number of required sensors, associated sensor types and sensor locations of sensors disposed within a living area of a senior according to a sensor floorplan.
4. The system of claim 1 , wherein at least one of the sensors is an Internet of Things (IOT) sensor device.
5. The system of claim 1 , wherein the sensor data is selected from:
a voice assistant appliance;
a video assistant appliance;
a smart phone;
a tablet computer;
a smart watch;
a smart appliance;
a personal computer; or
a home monitoring system.
6. A computer implemented method for providing in-home care for seniors, comprising:
generating a normalized feature vector from a set of sensor outputs from sensors disposed within a living area of a senior according to a sensor floorplan, including identifying a first relationship between symptoms of depression and events, wherein the events are measurable physical or mental features associated with at least one of the symptoms of depression, identifying a second relationship between the events and the set of sensor outputs based at least in part on the locations of each sensor and a sensor type of each sensor; and identifying a third relationship of sensor transformations required to transform sensor data into a format indicative of events, wherein the first relationship, the second relationship, and the third relationship is used to generate the normalized feature vector;
inputting the normalized feature vector to a logistic regression classifier of a machine learning model, wherein the logistic regression classifier is trained to determine thresholds for identifying depression based on a training data set of a set of seniors; and
determining a likelihood that the senior has depression.
7. A computer implemented method for providing in-home care for seniors, comprising:
mapping a set of sensor outputs from sensors disposed within a living area of a senior to events associated with features of symptoms of depression for the senior to generate an event vector for the senior identifying a first relationship between symptoms of depression and events, wherein the events are measurable physical or mental features associated with at least one of the symptoms of depression, identifying a second relationship between the events and the set of sensor outputs based at least in part on the locations of each sensor and a sensor type of each sensor; and identifying a third relationship of sensor transformations required to transform sensor data into a format indicative of events;
generating a normalized feature vector from the sensor outputs based on the first relationship, the second relationship, and the third relationship;
classifying the event vector into a likelihood that the senior has depression using a classifier trained to determine thresholds for identifying depression based on a training data set of a set of seniors; and
generating an output indicative of a risk a senior is depressed.
8. The method of claim 7 , wherein the sensors comprise sensors of user devices of the senior and Internet of Thing (IOT) devices.