Mobile Device Anomaly Detection via Classifier Confirmation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a need for effective behavioral monitoring and assessment of vulnerable individuals, such as the elderly, disabled, and those recovering from surgery or trauma, to timely detect medical anomalies like disorientation, seizures, or physical injuries, especially when they are not under supervised care, as timely detection is critical to prevent injuries or fatalities.
Innovation Solution
A computer-implemented method using sensor data from mobile devices to predict user states, transmitting confirmation requests to users, and sending notifications to secondary users if the user fails to respond or confirms the predicted state, employing classifiers trained on sensor data and geographic conditions to identify states like falls, disorientation, or seizures, with features like GPS and accelerometer data analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If active behavioral monitoring is implemented using mobile devices, then detection accuracy of medical anomalies is improved, but device complexity and power consumption increase
Solution Approach 1:
The monitoring system is divided into multiple independent components: sensor data collection module, classifier engine for anomaly detection, confirmation request module, and notification system. Each component operates independently but coordinates through standardized interfaces, reducing overall system complexity while maintaining high detection accuracy through specialized function distribution.
Solution Approach 2:
A confirmation request mechanism serves as an intermediary between the classifier engine and the notification system. When the classifier detects a potential anomaly, it first issues a confirmation request to the user's device before triggering alerts to emergency contacts. This intermediary step filters false positives and reduces unnecessary notifications while maintaining high true positive detection rates.
2Loss of time
If continuous monitoring is performed to ensure timely detection, then response time is reduced, but energy consumption increases
Solution Approach 1:
The system implements periodic monitoring at optimized intervals rather than continuous monitoring. The classifier engine processes sensor data at scheduled intervals, and the confirmation request mechanism is triggered periodically when anomalies are detected. This periodic approach maintains timely detection capability while significantly reducing power consumption compared to continuous operation.
Solution Approach 2:
The system uses the mobile device's existing sensors and processing capabilities to perform self-monitoring without requiring additional power-intensive hardware. The classifier engine leverages the device's own processor and memory resources, and the confirmation requests utilize existing communication channels, allowing continuous monitoring functionality with minimal additional energy consumption.
3Reliability
If confirmation requests are sent to users before notifications, then false alarms are reduced, but detection time is extended
Solution Approach 1:
The system performs preliminary classification of sensor data before issuing full anomaly notifications. The classifier engine pre-processes sensor readings and identifies potential anomalies, then sends confirmation requests to verify these preliminary findings. This preliminary action filters out false positives early in the process, improving overall alarm accuracy while keeping the additional time required minimal through efficient processing.
Solution Approach 2:
The confirmation request mechanism implements a feedback loop where the user's response to the confirmation request validates or refutes the classifier's preliminary anomaly detection. This feedback immediately confirms true positives and dismisses false positives, improving alarm accuracy. The system optimizes this feedback process by setting reasonable timeout periods and using push notifications for rapid user response, minimizing the time extension while maximizing reliability.
Data Source
AI summary
A computer-implemented method is provided including receiving sensor data from a mobile device corresponding to a first user. A user state of the first user is predicted based on the sensor data. A request is transmitted to the first user to confirm the predicted user state, and a notification is transmitted regarding the predicted user state to a second user responsive to the first user's confirmation of the predicted user state or the first user's failure to respond to the request. A computing system for monitoring and reporting activity of a mobile device is also provided.