Autonomous Vehicle Route Control with Asymmetric Potential Fields
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle travel route generation methods using Support Vector Machines often result in wavy routes due to obstacles, causing discomfort to occupants, especially when there are parked vehicles on one side of the road.
Innovation Solution
A travel control method and device that detect travelable road areas, generate a potential field with differing left and right boundary line potential values, compare and correct the potential values to align with pre-set lateral positions, and regenerate the travel route to ensure smooth autonomous vehicle travel.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the travel route is generated to pass through the center of the road area using Support Vector Machine, then the route avoids obstacles on one side, but the route becomes wavy and causes discomfort to occupants
Solution Approach 1:
The patent applies local quality by differentiating the potential values assigned to left and right boundary lines based on local obstacle conditions. When obstacles are detected on one side, the potential value for that boundary line is adjusted, creating an asymmetric potential field that locally modifies the travel route to maintain smoothness while avoiding obstacles. This resolves the contradiction by allowing obstacle avoidance only where necessary rather than forcing the route through the absolute center.
Solution Approach 2:
The patent changes the potential value parameters of boundary lines dynamically based on obstacle detection. By adjusting the potential values of left and right boundary lines according to the presence and position of obstacles, the system generates smooth travel routes that adapt to local conditions. This parameter adjustment resolves the contradiction between obstacle avoidance and route smoothness.
2Ease of operation
If the travel route is generated using a symmetric potential field with equal boundary values, then the route remains smooth, but the route cannot adapt to obstacles on one side of the road
Solution Approach 1:
The patent transforms the static symmetric potential field into a dynamic asymmetric potential field that adapts to obstacle conditions. The potential values of boundary lines are adjusted dynamically based on detected obstacles, allowing the system to maintain route smoothness while adapting to one-sided obstacle conditions. This dynamic adjustment resolves the contradiction between smoothness and adaptability.
Solution Approach 2:
The patent introduces asymmetry into the potential field by assigning different potential values to left and right boundary lines when obstacles are detected on one side. This asymmetric configuration allows the travel route to naturally deviate from the center toward the obstacle-free side while maintaining smoothness, resolving the contradiction between symmetric smoothness and asymmetric adaptability.
3Device complexity
If the Support Vector Machine sets the feature point closest to the identification plane as the Support Vector, then the calculation is simplified, but the generated route does not consider pre-set lateral positions and causes occupant discomfort
Solution Approach 1:
The patent introduces a potential field as an intermediary between the Support Vector Machine calculation and the final travel route generation. The potential field, with its asymmetric boundary values, acts as a mediator that guides the route generation process to respect pre-set lateral positions while maintaining calculation efficiency. This intermediary resolves the contradiction by adding guidance functionality without significantly increasing computational complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The travel control method comprises: detecting a travelable road area (TA) of the road area in which the subject vehicle (V) can travel; generating a potential field in a space of the travelable road area in which a potential value (+P) of a left-side boundary line and a potential value (-P) of a right-side boundary line are set to different values from each other; calculating a travelable road area width (W) by applying the Potential Method; comparing the difference between a lateral position of a travel route set on the basis of the calculated travelable road area width and a lateral position of the travel route set in advance in the travelable road area on the basis of the left-side boundary line or the right-side boundary line; correcting the potential value when there is a difference; regenerating the potential field set to the corrected potential value; generating the travel route (TR) applying the Potential Method to the regenerated potential field; executing autonomous travel control of the subject vehicle according to the generated travel route.