Skin Conductance Signal Processing for Sleep State Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to effectively reduce sleep onset time and optimize sleep quality and duration, particularly after a busy or emotionally challenging day, as they do not adequately account for the user's emotional state in providing sleep preparatory advice.
Innovation Solution
A signal processing device that classifies users into 'unhealthy tired' or 'healthy tired' states based on skin conductance data, using an input unit, segmentation, peak detection, and analysis to provide personalized sleep preparatory advice, such as wind-down recommendations, to facilitate quicker and more restful sleep.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing systems provide general sleep advice, then implementation is simple, but sleep onset time is not effectively reduced and sleep quality is not optimized
Solution Approach 1:
The system changes the parameter of emotional state assessment by using skin conductance data to distinguish between 'unhealthy tired' and 'healthy tired' states. This parameter change enables personalized sleep advice that adapts to the user's actual physiological condition, thereby improving sleep effectiveness without requiring complex invasive monitoring equipment.
Solution Approach 2:
The patent replaces complex mechanical or invasive monitoring systems with a simple skin conductance sensor that can be easily integrated into wearable devices. This substitution maintains high reliability for sleep effectiveness while keeping the system simple and non-intrusive.
2Adaptability or versatility
If skin conductance monitoring is implemented to assess emotional state, then personalized sleep advice can be provided, but measurement and processing complexity increases
Solution Approach 1:
The skin conductance signal is segmented into discrete events (peaks) that represent emotional responses. By segmenting the continuous signal into meaningful discrete units, the system achieves high adaptability for personalized assessment while keeping processing manageable through event-based analysis rather than continuous complex computation.
Solution Approach 2:
The system creates a simplified representation (copy) of the emotional state based on skin conductance patterns. Instead of processing all raw physiological data, it extracts key features (peak detection, rising edge amplitudes) that capture the essential emotional information, enabling personalization without full-signal processing complexity.
3Reliability
If comprehensive sleep analysis is performed without considering emotional state, then processing is faster, but sleep onset time and sleep quality cannot be optimized
Solution Approach 1:
The system performs preliminary assessment of the user's emotional state during the day using skin conductance monitoring. By determining whether the user is in an 'unhealthy tired' or 'healthy tired' state before bedtime, the system can provide targeted sleep advice in advance, which directly addresses the factors causing delayed sleep onset and poor sleep quality without requiring extended analysis time at bedtime.
Data Source
AI summary
The present invention relates to a system for improving sleep effectiveness of a user. The present invention further relates to a signal processing device (10) for processing skin conductance data of a user, the device comprising: an input unit (11) for receiving a skin conductance data signal indicative of a skin conductance of the user; a segmentation unit (12) for segmenting the skin conductance data signal into a plurality of epochs; a peak detection 5 unit (13) for detecting peaks in the skin conductance data signal; a calculation unit (14) for calculating a sum of rising edge amplitudes of the detected peaks per epoch; an analysis unit (15) configured to classify the user into an emotional state based on a transient behavior of said sums of rising edge amplitudes per epoch during the course of a day, wherein the analysis unit is configured to classify the user into an unhealthy tired state when the sum of 10 rising edge amplitudes per epoch increases during the course of the day; and/or to classify the user into a healthy tired state when the sum of rising edge amplitudes per epoch decreases during the course of the day; and an output unit configured to output an output signal indicative of said emotional state. The present invention further relates to a corresponding method and computer program implementing such a method.