ENVIRONMENTAL SENSOR-BASED COGNITIVE ASSESSMENT

• US 20160314255A1
• Filed: 04/21/2016
• Published: 10/27/2016
• Est. Priority Date: 04/21/2015
• Status: Active Grant

• Current Assignee: Washington State University
• Original Assignee: Washington State University
• Inventors: Edgecombe, Maureen E., Dawadi, Prafulla N., Cook, Diane J.
• Granted Patent: US 10,896,756 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/254 Extract, transform and load [ETL] procedures, e.g. ETL data flows in data warehouses
• CPC Class Codes: G06F 16/906 Clustering; Classification
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 20/10 using kernel methods, e.g. support vector machines [SVM]
• CPC Class Codes: G06N 20/20 Ensemble learning
• CPC Class Codes: G16H 40/67 for remote operation
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16Z 99/00 Subject matter not provided for in other main groups of this subclass

Abstract

Methods, systems, and techniques for facilitating cognitive assessment are provided. Example embodiments provide a Cognitive Assessment Facilitator System CAFS, which facilitates the gathering and prediction of cognitive assessment of individuals using machine learning and sensors placed in the home of a resident. These predictive assessments can then be used by a clinician to further diagnose and/or provide health intervention. In one embodiment, the CAFS comprises a sensor input module, a machine learning engine (or algorithm as part of another component), a CAAB tool, and activity curve change engine (activity tools), and a reporting module 308. These components cooperate to process and transform smart home based sensor data into activity performance features and statistical activity features which are then processing through a machine learning engine to predict clinical cognitive assessment values.

Claims

1. A method in a computing system for facilitating a cognitive assessment of an individual, comprising;

receiving from a plurality of smart sensors installed in a residence, raw output data of sensor events, the sensor events corresponding to a plurality of activities engaged in by the individual in the residence, the raw output data of each sensor event including at least a date, time, sensor identifier, and sensor value;

using automated activity recognition logic, automatically labeling the raw output data of each sensor event with an activity label that corresponds to a detected activity of the individual, the automatic labeling transforming the data into measurements of activities of daily living;

extracting activity performance features from the activity labeled sensor data; and

submitting indications of the extracted activity performance features to a machine learning engine to predict a clinical cognitive assessment based upon the sensor event data.

2. The method of claim 1 wherein the submitting indications of the extracted activity performance features include extracting statistical features from the extracted activity performance features to submit to the machine learning engine.

3. The method of claim 1, further comprising forwarding the predicted clinical cognitive assessment to a clinician for further diagnosis of cognitive state.

4. The method of claim 1 wherein the plurality of smart sensors comprise one or more motion sensors, light sensors, door sensors, or temperature sensors, wherein each sensor is associated with an identifier that identifies a location in the residence.

5. The method of claim 1 wherein the automatically labeling of the raw output data of each sensor event with an activity label includes labeling the data with activities indicating two or more of mobility, sleep, or activities of daily living.

6. The method of claim 5 wherein the activities of daily living comprise cooking, eating, relaxing, personal hygiene, or leaving home.

7. The method of claim 5 wherein the sleep label includes sleeping and a bed to toilet transition.

8. The method of claim 1 wherein the extracting activity performance features includes extracting time-based features and sensor-based features.

9. The method of claim 8 wherein the time-based features include an indication of duration of an associated activity.

10. The method of claim 8 wherein the sensor-based features include an indication of number of corresponding sensor events.

11. The method of claim 1 wherein extracting activity performance features further comprises extracting statistical activity features from the extracted activity performance features and the extracted statistical activity features include one or more of measures of variance, autocorrelation, skewness, kurtosis, or change.

12. The method of claim 1 wherein the sensor data is received continuously and wherein the activity recognition logic annotates and transforms the data by viewing the data in continuously moving evaluation windows over the data such that each evaluation window provides context for evaluating a newly received input of sensor data.

13. The method of claim 12 wherein the sensor data is received in near real time.

14. The method of claim 1 wherein indications of the extracted activity performance features are compared with actual clinical assessments to train the machine learning engine such that clinical data of the resident is used as a baseline for predictive assessment.

15. The method of claim 1 wherein the machine learning engine is a support vector machine.

16. A cognitive facilitator assessment computing system comprising;

a memory;

a computing processor;

a machine learning engine; and

clinical assessment activity behavior tool stored in the memory and further comprising;

sensor input logic configured, when executed, to provide live smart-home based sensor data corresponding to a resident in a residence installed with a plurality of smart-home based sensors;

an activity recognizer configured, when executed, to receive the sensor data real time and to transform the data with activity labels;

an activity performance features extractor configured, when executed, to extract activity performance features from the activity labeled sensor data, the activities corresponding to activities of daily living and/or mobility;

a statistical performance feature extractor configured, when executed, to extract statistical features of the extracted activity performance features; and

a cognitive assessment predictor configured, when executed, to receive the extracted performance features and to submit them to the machine learning engine to provide a clinical cognitive assessment.

17. The cognitive facilitator assessment computing system of claim 16 wherein the cognitive assessment predictor is further configured, when executed, to train the machine learning engine by initially submitting extracted statistical features of the extracted activity performance features with actual clinical assessments.

18. The cognitive facilitator assessment computing system of claim 16 wherein the extracted statistical features of the extracted activity performance features comprises at least one of variance, autocorrelation, skewness, kurtosis, or change.

19. The cognitive facilitator assessment computing system of claim 16 wherein the extracted activity performance features include time-based features and sensor-based features.

20. A non-transitory computer-readable memory medium containing contents for instructing a computer process to facilitate a cognitive assessment of a resident of residence with a plurality of smart sensors installed by performing a method comprising:

receiving from the plurality of smart sensors, continuous raw output data of sensor events, the sensor events corresponding to a plurality of activities engaged in by the individual in the residence, the raw output data of each sensor event including at least a date, time, sensor identifier, and sensor value;

automatically labeling the raw output data of each sensor event with an activity label that corresponds to a detected activity of the individual, the automatic labeling transforming the data into measurements of activities of daily living;

extracting activity performance features from the activity labeled sensor data;

extracting statistical activity performance features from the extracted activity performance features; and

submitting indications of the extracted statistical activity performance features to a machine learning engine to predict a clinical cognitive assessment based upon the continuous sensor event data.