Reactive Path Planning With Quintic Polynomials for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing path planning techniques for autonomous vehicles often fail to produce a complete solution within the given time, leading to inefficiencies and potential safety issues.
Innovation Solution
A path planning system that includes a reactive path planner based on constrained quintic polynomials, which generates alternative paths parallel to a nominal path within a predefined lateral distance, and evaluates these paths using a cost function to select the most feasible and efficient path.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing path planning techniques are used, then path planning can be performed, but complete solution cannot be found within the given time
Solution Approach 1:
The path planning problem is segmented into generating N alternative paths based on a nominal path, where each alternative path is displaced within a predefined lateral distance. This segmentation allows the system to explore multiple feasible paths without exhaustively searching the entire configuration space, thereby finding a complete solution within limited computation time.
Solution Approach 2:
The system uses constrained quintic polynomials to parameterize the alternative paths, changing the mathematical representation from traditional methods. By parameterizing paths with polynomial functions that incorporate vehicle dynamic and mechanical constraints, the system efficiently generates feasible paths while ensuring completeness within the given time frame.
2Reliability
If path planning constraints are applied to ensure safety and comfort, then passenger safety is improved, but computation complexity increases
Solution Approach 1:
Vehicle dynamic and mechanical constraints are incorporated directly into the polynomial parameters. The constrained quintic polynomials include parameters that represent maximum curvature, lateral acceleration, and steering rate limits, allowing safety and comfort constraints to be enforced through parameter boundaries rather than complex iterative checks.
Solution Approach 2:
Traditional mechanical constraint checking methods are replaced with mathematical polynomial constraints. Instead of simulating vehicle dynamics and checking constraints at each point along the path, the system uses polynomial equations that inherently satisfy the constraints, reducing computational complexity while maintaining safety and comfort assurances.
3Adaptability or versatility
If multiple alternative paths are generated, then path evaluation options increase, but computation time increases
Solution Approach 1:
The path space is segmented into N alternative paths that are parallel to the nominal path and displaced within a predefined lateral distance. This segmentation provides multiple evaluation options for adaptability while limiting the search space to only those paths that are likely to be feasible, thereby controlling computation time.
Solution Approach 2:
Instead of generating all possible paths through the environment, the system generates a partial set of N alternative paths that are sufficient for evaluating different scenarios. This partial action approach provides adequate path selection flexibility for adaptive decision-making without the excessive computation time required for exhaustive path generation.
Data Source
AI summary
A reactive path planner for vehicles is disclosed. The reactive path planner, based on a nominal path from A to B, generates alternative parallel paths which are displaced by a predefined lateral distance from the nominal path using constrained quintic polynomials. Constraints are imposed on the alternative paths to ensure safety and comfort based on dynamic and mechanical feasibilities of the vehicle. A cost function is applied to the nominal path and alternative paths select the path with the least cost. The reactive path planner achieves a complete path planning solution with high computational efficiency, even in highly cluttered and dynamic environments.