Driver Anomaly Detection Using Speed-Dependent Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mechanisms for detecting vehicle driver states based on steering activities fail to account for different driver states reflected by steering activities at various frequencies, particularly missing indicators of fatigue or drowsiness.
Innovation Solution
A vehicle system that collects steering angle and yaw-rate data, applies speed-dependent weighting, calculates recursive variances, and uses band-pass filtering to detect anomalous driving conditions by determining if anomaly indexes exceed predetermined thresholds, enabling the identification of driver drowsiness without additional road information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional steering activity detection is used, then basic driver state monitoring is achieved, but detection precision for different driver states (especially drowsiness) deteriorates due to inability to distinguish frequency-specific patterns
Solution Approach 1:
The patent segments the steering activity signal into different frequency components using spectral analysis. By dividing the continuous steering signal into distinct frequency bands, the system can analyze specific frequency ranges that correspond to different driver states (e.g., drowsiness vs. normal driving), thereby improving detection precision without requiring complete redesign of the monitoring system
Solution Approach 2:
The patent implements dynamic adjustment of detection parameters based on vehicle operating conditions. The system adapts its analysis by considering vehicle speed and other contextual factors to dynamically weight and interpret steering frequency patterns, allowing the detection system to remain effective across varying driving scenarios without fixed rigid thresholds
2Reliability
If frequency-based steering analysis is implemented, then detection of drowsiness and fatigue improves, but computational requirements and processing time increase
Solution Approach 1:
The patent extracts only the relevant frequency components from the complete steering signal spectrum. Rather than analyzing all frequency data, the system identifies and focuses on specific frequency bands that are most indicative of drowsiness and fatigue, discarding or minimizing processing of irrelevant frequency ranges. This extraction approach maintains high detection reliability while reducing overall computational burden
Solution Approach 2:
The patent applies partial action by performing spectral analysis only on selected segments of steering data or only when certain conditions are met (e.g., during highway driving at steady speeds). This selective application of complex frequency analysis reduces average computational energy consumption while maintaining detection reliability when it is most needed
3Adaptability or versatility
If multiple anomaly indexes are calculated, then comprehensive driver state coverage is achieved, but system complexity and threshold management difficulty increase
Solution Approach 1:
The patent designs the anomaly index calculation system to serve multiple detection purposes simultaneously. The same set of frequency-based anomaly indexes is used to detect various driver states including drowsiness, fatigue, and distraction, rather than having separate specialized indexes for each condition. This multi-functional approach achieves comprehensive driver state coverage while avoiding the complexity of managing entirely separate detection systems for each driver state
Data Source
AI summary
Vehicle trajectory data is obtained. A variance of the trajectory data is recursively determined. A speed-dependent weighting function is computed to obtain a speed-dependent weight. An anomaly index is determined based at least on part on application of the speed dependent weight to the recursively determined variance.