Driver Monitoring Gaze Criteria for Stopped-Vehicle Hazard Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driver monitoring systems fail to accurately determine whether a driver's visual recognition in the direction of travel is sufficient, especially in scenarios with stopped or decelerating surrounding vehicles, leading to potential collisions.
Innovation Solution
A driver monitoring device that acquires surrounding vehicle and road environment information, along with driver line-of-sight data, to adjust non-visual recognition determination criteria, such as gaze time or line-of-sight region, based on the number of stopped or decelerating vehicles, to suitably assess the driver's visual recognition and alert the driver when recognition is insufficient.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the same determination criterion is used for driver's visual recognition regardless of the number of stopped or decelerating surrounding vehicles, then the determination process remains simple, but the determination accuracy deteriorates and cannot suitably assess driver attention in different risk scenarios
Solution Approach 1:
The patent applies dynamics by making the determination criterion adjustable rather than fixed. The control device changes the non-visual recognition determination criterion (such as gaze time threshold or line-of-sight region size) based on the number of stopped or decelerating surrounding vehicles detected. This allows the system to adapt to different risk levels dynamically, improving determination accuracy without requiring a completely complex separate system for each scenario
Solution Approach 2:
The patent implements parameter changes by modifying specific parameters of the determination criterion based on surrounding vehicle conditions. When more stopped or decelerating vehicles are detected, the system adjusts parameters such as requiring longer gaze time or narrowing the acceptable line-of-sight region. This approach maintains relatively simple system structure while achieving accurate determination through parameter adjustment rather than structural complexity
2Reliability
If the non-visual recognition determination time is shortened to detect insufficient visual recognition more easily, then the sensitivity of detection improves, but the false alarm rate increases when the driver needs more time to recognize objects
Solution Approach 1:
The patent applies dynamics by making the non-visual recognition determination time adjustable based on surrounding vehicle conditions. When stopped or decelerating vehicles are detected, the system shortens the determination time threshold to increase detection sensitivity. When such vehicles are not present, the system extends the threshold to allow adequate recognition time, reducing false alarms. This dynamic adjustment resolves the contradiction between detection reliability and avoiding false positives
Solution Approach 2:
The patent implements parameter changes by modifying the time threshold parameter of the determination criterion based on the number of stopped or decelerating surrounding vehicles. The system changes this parameter dynamically: shorter thresholds when risk is high (more stopped vehicles), and longer thresholds when risk is low. This allows the system to maintain high reliability across different scenarios without fixed time loss
3Measurement precision
If the line-of-sight region is narrowed to focus on critical areas, then the precision of visual recognition assessment improves, but the coverage area for detecting potential hazards decreases
Solution Approach 1:
The patent applies dynamics by making the line-of-sight region size adjustable based on surrounding vehicle conditions. When stopped or decelerating vehicles are detected in specific locations, the system narrows the line-of-sight region to focus on critical areas where hazards are most likely, improving assessment precision. When such vehicles are not present, the system expands the region to maintain broader coverage. This dynamic adjustment resolves the contradiction between precision and coverage
Solution Approach 2:
The patent implements local quality by making different regions of the driver's field of view subject to different assessment criteria based on surrounding vehicle locations. When a stopped vehicle is detected, the system focuses the narrowed line-of-sight region assessment on the area adjacent to that vehicle where pedestrians or objects are most likely to emerge. This allows precise assessment in critical local areas while maintaining appropriate coverage through selective focus
Data Source
AI summary
A driver monitoring device has a processor which acquires surrounding vehicle information, surrounding road environment information and driver line-of-sight direction information, performs autonomous driving or driving assistance of a host vehicle based on the surrounding vehicle information and the surrounding road environment information, determines whether the driver's visual recognition in a direction of travel of the host vehicle is sufficient, alerts the driver when the driver's visual recognition in the direction of travel of the host vehicle is not sufficient, changes non-visual recognition determination criterion used for determining that the driver's visual recognition in the direction of travel of the host vehicle is not sufficient; and changes the non-visual recognition determination criterion so that the greater the number of the stopped or decelerating surrounding vehicles in the surrounding region, the easier it becomes to be determined that the driver's visual recognition is not sufficient.