Wearable Sleep Stage Detection Using Logistic Regression on Vital Sign Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current sleep stage detection methods, such as Polysomnography (PSG), are impractical for mass observations due to their immobility and high cost, and existing wearable sensors suffer from poor accuracy and susceptibility to sensor readings.
Innovation Solution
A method using logistic regression operations on vital sign features like heart rate and heart rate variability to detect sleep stages, which requires minimal computing resources and can be implemented in wearable devices, incorporating multiple logistic regression operations and machine learning techniques to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Polysomnography (PSG) with numerous sensors is used for sleep stage detection, then measurement precision is improved, but device complexity and cost increase significantly
Solution Approach 1:
The patent extracts and focuses on a specific subset of vital sign features (heart rate, pulse shape variability, and convoluted high frequency power of heart rate variability) from the comprehensive PSG sensor suite. By selecting only the most relevant features for sleep stage detection, the system achieves acceptable accuracy with significantly fewer sensors, directly resolving the contradiction between measurement precision and device complexity
Solution Approach 2:
The patent makes wearable sensors multi-functional by using them to detect multiple sleep stages (REM and NREM) through a single integrated system that processes multiple vital sign features. This allows one wearable device to perform comprehensive sleep stage detection without requiring separate specialized sensors for each function, reducing overall device complexity while maintaining detection accuracy
2Measurement precision
If Polysomnography (PSG) with numerous sensors is used for sleep stage detection, then measurement precision is improved, but ease of operation deteriorates due to immobility
Solution Approach 1:
The patent employs inexpensive wearable sensors that can be easily disposed of or replaced, contrasting with the expensive, complex PSG equipment. These wearable sensors provide sufficient accuracy for sleep stage detection while being portable and easy to operate, directly addressing the contradiction between measurement precision and ease of operation
Solution Approach 2:
The patent replaces the complex mechanical and electrical PSG system with a simplified wearable sensor system that uses photoplethysmography and other non-invasive techniques. This substitution maintains adequate measurement precision for sleep stage detection while dramatically improving portability and ease of operation
3Ease of operation
If existing wearable sensors are used for sleep stage detection, then ease of operation is improved, but measurement precision deteriorates due to poor accuracy
Solution Approach 1:
The patent performs preliminary processing and feature extraction on raw sensor data, calculating specific vital sign features (heart rate, pulse shape variability, convoluted high frequency power of heart rate variability) before sleep stage detection. This preliminary action improves measurement precision by preparing optimized input data for the detection algorithm, while maintaining the ease of operation provided by wearable sensors
Solution Approach 2:
The patent introduces an intermediary processing layer that transforms raw wearable sensor readings into meaningful sleep stage information through logistic regression operations. This intermediary layer bridges the gap between the limited capabilities of wearable sensors and the requirements for accurate sleep stage detection, improving precision without sacrificing wearability
4Ease of operation
If existing wearable sensors are used for sleep stage detection, then ease of operation is improved, but reliability deteriorates due to susceptibility to poor sensor readings
Solution Approach 1:
The patent employs beforehand cushioning by using logistic regression with data from multiple epochs (preceding, current, and succeeding epochs) to compensate for poor sensor readings. This approach provides a buffer against unreliable measurements by considering temporal context, thereby improving reliability while maintaining the ease of operation of wearable devices
Data Source
AI summary
A method of sleep stage detection using vital sign features derived from PPG signals. The method includes performing a logistic regression operation based on a machine learning classifier model to calculate an indication value for an intermediate epoch. The indication value is calculated based on the vital sign features for the intermediate epoch as well as those of the preceding and succeeding epochs. The method then detects the sleep stage of the corresponding intermediate epoch based on the indication value for the corresponding intermediate epoch.