Sleep State Classification Using Motion Sensor Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for tracking sleep state using wearable devices are limited in their ability to accurately classify and provide detailed information about sleep stages, often failing to distinguish between different sleep states and requiring complex processing and multiple sensors.
Innovation Solution
A wearable device equipped with motion and orientation sensors, such as accelerometers and gyroscopes, processes data to classify sleep states as awake, REM sleep, non-REM sleep stages, and filters the results to improve accuracy, using machine learning models and quality checks to ensure reliable data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complex processing and multiple sensors are used to accurately classify sleep stages, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent segments the sleep classification process into multiple hierarchical stages: first classifying broad sleep/wake states, then progressively refining to specific sleep stages (REM, NREM stages 1-3). This segmentation allows the system to achieve high measurement precision through systematic multi-stage analysis rather than attempting single-step classification, thereby managing processing complexity through structured decomposition of the classification task.
Solution Approach 2:
The patent applies preliminary filtering and quality checks to sensor data before classification. Motion artifacts are detected and filtered out in advance, and data quality is assessed before proceeding with sleep stage classification. This preliminary action ensures that only high-quality data undergoes complex processing, improving measurement precision while reducing unnecessary computational burden on low-quality data.
2Measurement precision
If multiple sensors and complex processing are used to distinguish different sleep states, then measurement precision is improved, but ease of operation deteriorates
Solution Approach 1:
The system automatically performs quality checks and motion artifact detection without requiring user intervention. The wearable device self-manages the classification process, automatically filtering motion artifacts and assessing data quality before generating sleep stage reports. This self-service approach maintains high measurement precision while preserving ease of operation, as users simply wear the device and receive results without needing to understand or configure the complex processing underlying the classification.
3Reliability
If sleep tracking session start and end are defined by rest and activity states, then reliability is improved, but measurement precision may worsen due to state transition ambiguity
Solution Approach 1:
The system uses feedback from motion sensors and classification results to continuously refine sleep state detection. During transitions between rest and activity states, the system monitors classification confidence and adjusts detection thresholds dynamically. This feedback mechanism allows the system to maintain reliable sleep tracking by confirming state transitions through multiple data points and classification iterations, thereby resolving ambiguity and improving both reliability and measurement precision at state transitions.
Data Source
AI summary
A wearable device including a motion-tracking sensor can be used for tracking sleep. The data from the motion-tracking sensor can be to estimate/classify the sleep state for multiple periods and/or to determine sleep intervals. In some examples, to improve performance, sleep state classification can be performed on data within a sleep tracking session. The start of the sleep tracking session can be defined by detecting a rest state and the end of the sleep tracking session can be defined by an activity state. In some examples, to improve performance, the classified sleep states for the multiple periods can be filtered and/or smoothed. In some examples, a signal quality check can be performed for the data from the motion-tracking sensor. In some examples, the classification of the sleep states and/or the display of the results of sleep tracking can be subject to passing one or more signal quality checks.