Personalized PERCLOS Thresholds for Driver Drowsiness Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing drowsiness detection systems face challenges in accurately detecting drowsiness due to personal differences in PERCLOS thresholds, leading to inconsistent detection accuracy when uniform threshold values are used.
Innovation Solution
A drowsiness detection device that adjusts the threshold value based on the initial ratio and number of deep blinks of the driver, using a camera to analyze eye closure patterns and set personalized thresholds for improved detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a uniform threshold value is used for PERCLOS detection, then the detection system is simple to implement, but detection accuracy deteriorates due to personal differences in blink patterns
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the PERCLOS threshold value based on individual driver characteristics. Instead of using a fixed uniform threshold, the system calculates personalized threshold values according to each driver's baseline blink patterns, thereby resolving the contradiction between implementation simplicity and detection accuracy.
Solution Approach 2:
The system implements self-service by automatically adapting to each driver's unique characteristics without requiring manual intervention. The threshold value is automatically calculated and adjusted based on the driver's own blink history, eliminating the need for manual calibration while maintaining high detection accuracy.
2Measurement precision
If the threshold value is made optionally changeable by the driver, then detection accuracy can be improved, but operation complexity increases requiring confirmation of past history
Solution Approach 1:
The system performs self-service by automatically calculating and setting personalized threshold values based on each driver's blink patterns. This eliminates the need for drivers to manually adjust thresholds or review their blink history, thereby maintaining detection accuracy while preserving operational simplicity.
Solution Approach 2:
The system performs preliminary action by pre-calculating personalized threshold values during the initial period of driver operation. This preliminary calculation is done automatically in the background, so when detection is needed, the personalized threshold is already ready without requiring any driver action or confirmation.
3Measurement precision
If personalized threshold values are set based on initial deep blink analysis, then detection accuracy improves for varying blink patterns, but device complexity increases
Solution Approach 1:
The system uses parameter changes by adjusting the threshold value based on measurable variations in blink patterns. By focusing on specific parameters like deep blink ratio and duration, the system achieves personalized detection without requiring complex multi-dimensional analysis, thus balancing accuracy with manageable system complexity.
Solution Approach 2:
The system applies taking out by extracting only the most relevant features from blink data - specifically the deep blink ratio and duration - to determine personalized thresholds. This selective extraction of critical parameters simplifies the overall system complexity while maintaining high detection accuracy through focused analysis.
Data Source
AI summary
A drowsiness detection device includes a PERCLOS calculation unit, a drowsiness detection unit, and a threshold value setting unit. The PERCLOS calculation unit calculates, based on an image imaged by a camera, a PERCLOS that is a ratio of a time period during which a driver closes an eye within a certain time period. The drowsiness detection unit detects drowsiness of the driver when the PERCLOS is equal to or larger than a threshold value. The threshold value setting unit sets the threshold value to be smaller than in a case that an initial ratio of deep blinking is equal to or larger than a % when an initial ratio of deep blinking first calculated from a start of driving is less than a %.