MCI Detection via Passive Sensor Data Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting Mild Cognitive Impairment (MCI) using sensors invade privacy and are prone to technical issues, and high-dimensional sensor data collected by non-intrusive sensors cannot be effectively processed by existing machine learning techniques.
Innovation Solution
A system and method that utilize non-intrusive sensors in a smart environment to continuously monitor activities, preprocess sensor data using Long Short-Term Memory (LSTM) to correct corrupt data, generate semantic vector representations, optimize data size with autoencoders, and estimate dynamic thresholds to detect behavior deviations indicative of MCI.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If cameras or sensors are used to monitor routine activities of subjects, then detection capability of MCI progression is improved, but privacy invasion and technical reliability deteriorate
Solution Approach 1:
The patent introduces an intermediary processing layer (LSTM-based prediction model and autoencoder) that mediates between raw sensor data and detection outcomes. This intermediary system corrects corrupt sensor data using predictive models and optimizes data representation, thereby improving detection capability while reducing direct reliance on potentially unreliable individual sensors and minimizing privacy invasion through aggregated analysis.
2Object-affected harmful factors
If non-intrusive sensors are used to monitor Activities of Daily Living, then privacy invasion is reduced, but data dimensionality and processing complexity increase
Solution Approach 1:
The patent transforms the parameter representation of sensor data by converting high-dimensional raw sensor readings into optimized vector representations through autoencoders. This parameter transformation reduces data dimensionality while preserving essential behavioral patterns, thereby reducing processing complexity without compromising detection accuracy and maintaining the non-intrusive nature of the sensors.
3Measurement precision
If high-dimensional sensor data is collected from multiple sensors, then measurement completeness is improved, but data processing capability deteriorates
Solution Approach 1:
The patent extracts essential features from high-dimensional sensor data by using LSTM networks to capture temporal patterns and autoencoders to identify core behavioral representations. This extraction process separates essential measurement information from redundant data dimensions, maintaining measurement completeness while significantly improving data processing capability through dimensionality reduction.
4Measurement precision
If sensor data is continuously monitored over time, then detection accuracy of MCI progression is improved, but false alarms and computational overhead increase
Solution Approach 1:
The patent implements feedback mechanisms where LSTM networks learn from historical sensor data patterns to predict future behaviors and establish baseline norms for each subject. The system continuously compares actual sensor readings against these learned patterns, adjusting detection thresholds dynamically. This feedback approach improves detection accuracy for true MCI progression while reducing false alarms by distinguishing actual deviations from normal variations in behavior.
Data Source
AI summary
This disclosure relates generally to detection of mild cognitive impairments in subjects. The method and system proposed provides a continuous/seamless monitoring platform for MCI detection in subjects by continuously monitoring routine activities of subjects (Activities of Daily Living (ADL)) in a smart environment using plurality of passive, unobtrusive, binary, unobtrusive non-intrusive sensors embedded in living infrastructure. The proposed method and system detects symptoms of MCI at the onset of the disease, while also addressing issue of sensor failures that causes gaps in the data. The collected sensor data is pre-processed in several stages which includes which includes pre-processing of sensor data, behavior deviation detection, and abnormality detection and so on. Further, the disclosure also proposes an autoencoder based technique, to reduce the dimension of the data to find personalized deviations in behavior of a subject which is used to detect if a subject could be a potential case of MCI.