Autonomous Vehicle Motion Planning for Natural Tight Turns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles often make turns that feel unnatural to human passengers by strictly following nominal pathways, which can result in over-broad or wide turns deviating from traditional human driving behavior.
Innovation Solution
A computer-implemented method that projects a candidate motion plan onto a nominal pathway to determine a projected distance, using a reward function positively correlated to the projected distance, allowing the autonomous vehicle to make tighter turns that feel more natural, optimizing the total cost based on cost and reward functions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the autonomous vehicle strictly follows the nominal pathway, then the vehicle maintains accurate path tracking, but the turns become over-broad and feel unnatural to passengers
Solution Approach 1:
Instead of directly following the nominal pathway centerline, the system inverts the approach by targeting a offset position (e.g., 0.5-2 meters from the centerline) as the new reference trajectory. This inversion allows the vehicle to make tighter, more natural turns while still maintaining accurate path tracking relative to the modified reference, resolving the contradiction between path accuracy and natural turning behavior
Solution Approach 2:
The system changes the parameter of the reference trajectory by introducing an offset distance from the nominal pathway centerline. By adjusting this offset parameter dynamically, the vehicle can achieve tighter turn radii that feel more natural to passengers while maintaining controlled path tracking accuracy through the modified reference frame
2Ease of operation
If the autonomous vehicle makes tighter turns, then passenger comfort and natural feeling improve, but the projected distance along the nominal pathway decreases
Solution Approach 1:
The system performs preliminary action by pre-calculating an offset reference trajectory that anticipates the need for tighter turns. By preparing this modified reference path in advance, the vehicle can execute natural-looking turns without sacrificing overall driving efficiency, as the offset trajectory is designed to maintain optimal progress toward the destination
3Device complexity
If the autonomous vehicle follows the nominal pathway centerline, then the motion plan is simple to compute, but the turn width is excessive and inefficient
Solution Approach 1:
The system applies parameter changes by modifying the reference trajectory parameter (introducing an offset from the centerline) rather than changing the fundamental motion planning algorithm. This approach maintains computational simplicity while achieving tighter, more efficient turns, as the same planning framework works with the offset parameter to produce improved results
Data Source
AI summary
The present disclosure provides systems and methods that control the motion of an autonomous vehicle by rewarding or otherwise encouraging progress toward a goal, rather than simply rewarding distance travelled. In particular, the systems and methods of the present disclosure can project a candidate motion plan that describes a proposed motion path for the autonomous vehicle onto a nominal pathway to determine a projected distance associated with the candidate motion plan. The systems and methods of the present disclosure can use the projected distance to evaluate a reward function that provides a reward that is positively correlated to the magnitude of the projected distance. The motion of the vehicle can be controlled based on the reward value provided by the reward function. For example, the candidate motion plan can be selected for implementation or revised based at least in part on the determined reward value.