Autonomous Vehicle Gap Planning With Reduced Horizon Path Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in quickly identifying lane gaps due to computationally expensive path planning processes, which may not complete before encountering obstructions, necessitating an improved system for efficient gap planning.
Innovation Solution
A computer-implemented method utilizing a gap planning module to identify lane gaps with a longer time horizon and generate planning constraints, separate from the path planning module, which uses a shorter time horizon to generate paths, thereby reducing computational load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the planning horizon is extended to successfully identify a gap, then the reliability of gap identification is improved, but the processing power requirement increases exponentially
Solution Approach 1:
The patent divides the path planning function into two separate modules: a gap planning module that identifies gaps using a first time horizon, and a path planning module that generates paths using a second time horizon. This segmentation allows gap identification to be performed with reduced computational requirements while maintaining reliability, as the gap planning module operates independently before the computationally intensive path planning begins.
Solution Approach 2:
The gap planning module performs preliminary gap identification before the full path planning process. By pre-identifying suitable gaps in adjacent lanes using a shorter first time horizon, the system prepares constraint information in advance that guides the subsequent path planning module, avoiding the need for the full path planning process to search the entire planning horizon for gap information.
2Reliability
If the planning horizon is extended to identify gaps, then the completeness of path planning is improved, but the computation time increases
Solution Approach 1:
The patent segments the planning horizon into two distinct time horizons: a first time horizon for gap planning and a second time horizon for path planning. This allows the system to perform gap identification quickly with a shorter horizon while maintaining overall planning completeness through the subsequent path planning stage that operates with a longer horizon if needed.
Solution Approach 2:
By performing gap identification as a preliminary action before full path planning, the system reduces the effective search space for the path planning module. The gap planning module quickly identifies candidate gaps and generates constraint information that guides the path planning module, significantly reducing the computation time required for complete path planning.
3Device complexity
If a single path planning module is used, then the system complexity is reduced, but the responsiveness to lane obstructions deteriorates
Solution Approach 1:
The patent introduces a segmented architecture with a gap planning module and a path planning module. The gap planning module is optimized for rapid gap identification and can operate with a shorter time horizon, enabling fast response to lane obstructions. While this increases system complexity compared to a single module, the modular design allows each component to be optimized for its specific function, improving overall responsiveness.
Solution Approach 2:
The gap planning module performs preliminary gap identification that can be completed quickly, providing immediate information about suitable lane changes when obstructions are detected. This preliminary action enables the system to respond faster to lane obstructions by having gap information ready before the full path planning process begins, improving lane change responsiveness despite the increased system complexity.
Data Source
AI summary
A computer-implemented method for controlling an autonomous vehicle is provided. The computer-implemented method includes identifying, using a gap planning module, based on a first set of perception data using a first time horizon, a gap an adjacent lane to a lane of travel of the autonomous vehicle, the adjacent lane suitable for entry by the autonomous vehicle. The computer-implemented method further includes generating, using the gap planning module, one or more planning constraints based on the identified gap. The computer-implemented method further includes generating, using a path planning module, a path for the autonomous vehicle based on a second set of perception data, a second time horizon, and the one or more planning constraints. The computer-implemented method further includes controlling the autonomous vehicle based on the generated path.