Path Planning Correction Using Lane Line Parallelism Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing path planning systems for vehicles face accuracy issues when the relationship between left and right lane lines changes, such as at intersections or where the number of lanes increases or decreases, leading to decreased accuracy in path planning corrections.
Innovation Solution
A path planning apparatus using a first deep learning model generates a path plan from a vehicle-mounted camera image, with a second deep learning model recognizing lane lines and determining if they satisfy a parallel condition; if not, the path plan is not corrected, employing an allowable angle and lane width threshold to prevent inaccurate corrections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If path planning correction is performed using lane line recognition results, then path planning accuracy is improved, but accuracy decreases at locations where lane line relationships change
Solution Approach 1:
The system performs preliminary detection of lane line relationships before correction. By detecting whether the angle between left and right lane lines exceeds the allowable threshold in advance, the system prevents incorrect corrections from being applied at intersections and branch roads, thereby maintaining correction reliability while preserving accuracy improvements.
Solution Approach 2:
The system incorporates feedback by continuously monitoring the angle between recognized lane lines and comparing it against the allowable threshold. When the feedback indicates that lane lines do not satisfy the parallel condition (angle exceeds threshold), the correction process is halted, preventing accuracy degradation at complex road locations.
2Manufacturing precision
If lane line recognition is used to correct path plans, then path planning precision is improved, but errors increase when lane lines are not parallel
Solution Approach 1:
The system performs preliminary angle calculation between lane lines before applying corrections. By calculating the angle in advance and comparing it with the allowable threshold, the system prevents precision errors from propagating when lane lines are not parallel, thus maintaining overall path planning precision.
Solution Approach 2:
The angle threshold condition acts as an intermediary filter between lane line recognition and path plan correction. This intermediary mechanism validates whether correction should be applied by checking if the angle between lane lines is within the allowable range, preventing direct transmission of recognition errors to the final path plan.
3Adaptability or versatility
If path plan correction is applied at all locations, then correction coverage is improved, but accuracy decreases at intersections and branch roads
Solution Approach 1:
The system applies different correction strategies to different locations based on local characteristics. By detecting whether the current location has lane lines satisfying the parallel condition (angle within threshold), the system enables correction only in appropriate local areas (straight roads) while avoiding correction in unsuitable areas (intersections, branch roads), thus maintaining overall accuracy while preserving correction coverage where beneficial.
Data Source
AI summary
The path planning apparatus according to an aspect of the present disclosure generates a path plan of a host vehicle using a first deep learning model from a captured image of a vehicle-mounted camera. The path planning apparatus recognizes left and right lane lines forming a travelling lane of a host vehicle from an image captured by a vehicle-mounted camera, and corrects a path plan using a recognition result of the lane lines. The path planning apparatus determines whether the left and right lane lines satisfy a preset parallel condition. The path planning apparatus does not correct the path plan using the recognition result of the lane lines, when the left and right lane lines are determined as not satisfying the parallel condition.