Driver Face Orientation Monitoring for Level 3 Takeover Alerts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driver monitoring systems struggle to accurately determine a driver's state during level 3 autonomous driving, leading to potential safety issues and decreased driver convenience due to unnecessary notifications.
Innovation Solution
A driver monitor that detects the orientation and eye opening of a driver's face using a camera, determining if the face inclination exceeds a threshold angle and time, and only notifies the driver to take over when necessary, based on these conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the driver monitoring system issues transition demand notifications frequently to ensure safety, then the safety level improves, but the driver convenience deteriorates due to unnecessary notifications
Solution Approach 1:
The system changes the parameter of notification timing based on detected driver state parameters (face orientation angle and duration). When the driver's face orientation exceeds a threshold angle for a threshold duration, the system transitions from no notification to issuing transition demand notification, optimizing the balance between safety and driver convenience by notifying only when necessary.
2Measurement precision
If the driver monitoring system requires the driver to be facing forward continuously, then the measurement precision of driver state improves, but the adaptability to autonomous driving conditions deteriorates
Solution Approach 1:
The system dynamically adjusts the driver state requirements based on the autonomous driving mode. During level 3 autonomous driving, the system allows temporary face orientation deviations beyond the threshold angle for durations exceeding the threshold time, recognizing that drivers may engage in secondary activities. The system issues transition demand notifications only when the driver state deviates too much for too long, providing adaptive monitoring that balances measurement precision with autonomous driving adaptability.
3Measurement precision
If the driver monitoring system uses strict face orientation thresholds to determine driver attention, then the measurement precision improves, but the loss of information about actual driver state increases
Solution Approach 1:
The system adds the time dimension to the face orientation measurement. Instead of using only the angle threshold, the system measures both the angle and the duration exceeding the threshold. This dual-parameter approach (angle + duration) provides a more comprehensive understanding of driver state, reducing information loss while maintaining measurement precision. The combination of spatial (angle) and temporal (duration) dimensions allows the system to distinguish between temporary glances and sustained inattention.
Data Source
AI summary
A driver monitor includes a processor configured to detect orientation of the face of a driver of a vehicle from a plurality of driver images generated by a camera mounted on the vehicle to take a picture of the driver, determine whether a duration during which the orientation of the driver's face inclines more than a predetermined angle in a yaw direction of the vehicle with respect to a reference direction is not less than a predetermined time threshold, when the vehicle is under autonomous driving control, notify a transition demand for taking over driving of the vehicle to the driver via a notification device when the duration is less than the predetermined time threshold, and omit to notify the transition demand to the driver when the duration is not less than the predetermined time threshold.