Non-obtrusive Emotional Loneliness Detection Using Motion Sensor Anomalies
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting emotional loneliness in elderly individuals are often intrusive, violating privacy and lack accuracy in identifying emotional isolation.
Innovation Solution
A non-obtrusive system and method using motion sensors to analyze room changes, living room stay, bedroom correlation, and outdoor activity anomalies, employing a local outlier factor model to determine emotional loneliness with reduced variance and false positives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensors are installed for tracking movement to detect loneliness, then detection accuracy is improved, but privacy is violated and the method becomes obtrusive
Solution Approach 1:
The patent extracts only the necessary movement data (room transitions, time spent in rooms, outdoor activity patterns) from the sensor data, rather than capturing comprehensive personal information. This selective extraction maintains detection accuracy while minimizing privacy intrusion by focusing solely on spatial-temporal patterns relevant to loneliness assessment.
Solution Approach 2:
The system uses motion sensors as intermediaries that indirectly measure loneliness through movement patterns rather than directly monitoring personal activities or communications. The sensors capture anonymized spatial data that serves as a mediator between the physical environment and emotional state assessment, preserving privacy while enabling detection.
2Object-affected harmful factors
If simple sensor tracking is used, then privacy is preserved, but detection accuracy is insufficient
Solution Approach 1:
The patent segments the detection approach into multiple independent analytical dimensions: room transition frequency, duration of stay in specific rooms, temporal patterns of outdoor activity, and correlation between bedroom and living room usage. Each segment contributes partially to the overall loneliness assessment, allowing accurate detection through aggregated patterns rather than relying on a single intrusive metric.
Solution Approach 2:
The system transitions from traditional single-dimensional loneliness assessment (e.g., self-report surveys) to multi-dimensional analysis by incorporating spatial (room locations), temporal (time patterns), and behavioral (movement frequency) dimensions. This dimensional expansion enables accurate detection using privacy-preserving motion data that captures complex loneliness indicators across multiple facets of daily life.
3Reliability
If multiple analysis methods are used to improve accuracy, then detection reliability is improved, but system complexity increases
Solution Approach 1:
The patent merges multiple analysis methods (room change analysis, living room stay analysis, bedroom correlation analysis, outdoor activity analysis) into a unified loneliness detection system. These methods are combined through a composite scoring mechanism that integrates results from all four analyses, achieving high reliability by leveraging complementary strengths of each method while operating as a coordinated system rather than separate independent tools.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Emotional loneliness is referred as the absence of an attachment figure in one's life and someone to turn to. The existing methods use installation of sensors for tracking the movement, behaviour and activity of the person, but most of the efforts are obtrusive in nature. A non-obtrusive method and system for detection of emotional loneliness of a person have been provided. The disclosure is utilizing multiple varied techniques to understand the emotional loneliness. The multiple techniques comprise room change movement anomalies, living room stay anomalies, correlating the living room stay with the bedroom stay and outdoor movement anomalies. The methodology also ensures reduced variance and false positives, as emotional loneliness is finally determined based on more than two positive of above methods. The detection of person's movement is done using a featured engineered dataset based on collection of raw time series data collected from a plurality of motion sensors.