Adaptive Lane Centering via Driver Behavior Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles often navigate differently than human drivers, leading to discomfort for passengers who prefer the behavior of a human driver, as they may hug the outer or inner edge of curves, whereas autonomous vehicles typically stay mid-way between edges.
Innovation Solution
A system and method where a processor learns a driver's behavior by measuring vehicle speed, lateral control, and acceleration/deceleration, constructing a knowledge matrix, and creating a behavior policy to adapt the vehicle's navigation to imitate the driver's behavior, including offline and online learning modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the autonomous vehicle navigates by maintaining itself midway between outer edge and inner edge of the curve, then the vehicle follows a predetermined autonomous behavior, but the driver comfort and preference are not satisfied
Solution Approach 1:
The system changes the behavioral parameters of the autonomous vehicle by learning from driver actions. The processor measures driver behavior parameters (steering angle, acceleration, braking) and adjusts the vehicle's navigation parameters to match the driver's preferred style, transforming the vehicle from fixed predetermined behavior to adaptive behavior that mimics the driver's patterns.
Solution Approach 2:
The system creates a behavioral copy of the driver's driving style. The processor records and analyzes driver actions during offline learning mode, then replicates these behaviors during autonomous operation. The vehicle essentially copies the driver's unique driving patterns, including how they navigate curves, change lanes, and respond to road conditions.
2Measurement precision
If the vehicle learns driver behavior through measuring speed, lateral control, and acceleration, then the driver preference adaptation is improved, but the data processing complexity and time increase
Solution Approach 1:
The system performs preliminary learning during offline mode before autonomous operation begins. The processor collects and analyzes driver behavior data in advance, building a comprehensive behavioral profile. This preliminary action ensures that when the vehicle operates autonomously, it already has the learned behavior patterns ready to apply, eliminating the need for real-time learning delays.
Solution Approach 2:
The learning process continues continuously across different driving sessions and conditions. The system accumulates behavioral data over multiple trips and road segments, continuously refining the driver profile. This continuous learning ensures the system maintains high measurement precision while spreading the processing workload over time rather than concentrating it all at once.
3Ease of operation
If the behavior policy is created based on driver behavior and safety threshold, then the driver comfort is improved, but the safety constraint compliance requires additional processing
Solution Approach 1:
The system preemptively applies safety constraints when creating the behavior policy. Before finalizing the learned driver behavior into actionable commands, the processor checks against predefined safety thresholds and limits. This preliminary anti-action prevents unsafe behaviors from being implemented, ensuring reliability is maintained while still allowing driver comfort preferences to shape the overall behavior pattern.
Solution Approach 2:
The system uses feedback loops to continuously monitor and adjust the balance between driver comfort and safety compliance. When the vehicle detects that learned driver behaviors approach safety boundaries, it provides feedback to modify the behavior policy accordingly. This feedback mechanism ensures that driver comfort preferences are honored while maintaining strict adherence to safety requirements.
Data Source
AI summary
A vehicle and a system and method of operating a vehicle. The system includes a processor. The processor learns a driver's behavior of a driver of the vehicle as the driver navigates a road segment, creates a behavior policy based on the driver's behavior and a threshold associated with the road segment, and controls the vehicle to navigate the road segment using the behavior policy.