Quadratic Path Planning in Frenet Space for Obstacle Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional path planning algorithms for unmanned vehicles, such as the RRT algorithm, result in paths with low quality due to random distribution of sampling points, leading to unsatisfactory smoothness and increased collision risks.
Innovation Solution
A path planning method using quadratic programming based on the mapping relationship between the Cartesian and Frenet coordinate systems to determine a target path, incorporating road and obstacle constraints, ensuring safety and smoothness through natural cubic spline curve interpolation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If path planning is performed in a 2D plane using conventional methods, then the planning process is simple and fast, but the path may collide with 3D obstacles such as bridges or overpasses
Solution Approach 1:
The patent transforms the path planning problem from 2D to 3D by constructing a three-dimensional configuration space that includes vertical height information. This allows the planner to detect and avoid collisions with bridges, overpasses, and other 3D obstacles that would be invisible in traditional 2D planning, directly resolving the contradiction between collision avoidance reliability and planning complexity.
Solution Approach 2:
The patent implements a hierarchical path planning architecture where a global 3D path planner operates at a higher level to generate an overall route, and a local 2D path planner operates within specific regions to refine the path. This nested structure allows the system to maintain simplicity in local planning while achieving 3D collision avoidance through global planning, resolving the complexity-reliability contradiction.
2Reliability
If path planning considers 3D spatial information including height, then collision avoidance with bridges improves, but the computational complexity and processing time increase
Solution Approach 1:
The patent divides the path planning process into multiple segments: a global 3D planning stage that handles overall route generation with bridge avoidance, and local 2D planning stages that handle detailed path refinement. This segmentation allows the system to perform 3D planning only where necessary while maintaining efficiency in other areas, resolving the time-complexity contradiction.
Solution Approach 2:
The patent performs preliminary 3D spatial analysis and obstacle detection before executing the main path planning algorithm. By pre-processing the environment data to identify bridges and overpasses in advance, the system reduces the computational burden during actual path planning, thereby reducing planning time while maintaining collision avoidance reliability.
3Productivity
If the robot follows a path generated without considering robot dimensions, then the planning process is faster, but the robot body may collide with obstacles
Solution Approach 1:
The patent applies different levels of detail to different parts of the path planning process. The global planner uses simplified representations for efficiency, while the local planner incorporates detailed robot dimensions and obstacle geometries in critical areas to prevent collisions. This local quality approach maintains overall efficiency while ensuring safety where needed.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A path planning method is provided, to implement planning of an obstacle avoidance path, so that quality of a planned path can be improved. The method includes: obtaining road information, an initial location of a target object, and a destination location (401); determining a Cartesian coordinate system, a Frenet coordinate system, and a mapping relationship between the Cartesian coordinate system and the Frenet coordinate system based on the road information; obtaining, based on the mapping relationship, a path quality indicator function based on the Frenet coordinate system, where the path quality indicator function is a quadratic function; determining a target path based on the Frenet coordinate system by using optimal quadratic programming and based on the path quality indicator function based on the Frenet coordinate system, the initial location of the target object, and the destination location; and converting the target path into the Cartesian coordinate system based on the mapping relationship (406), where the target path is used for path planning and control of the target object.