Motion Planning for Transportation Vehicles Using Action State Gridding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current motion planning techniques for transportation vehicles face challenges in achieving real-time capability due to high computational complexity, particularly in handling complex driving maneuvers like parking, where simplifications often compromise optimality and usability.
Innovation Solution
The method employs action and state gridding to optimize motion candidates, selecting final states in a discretized space, and using pre-calculated action discretization and heuristic rules to reduce computational complexity, enabling real-time motion planning by prioritizing promising candidates and simulating human-like driving behaviors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If comprehensive motion planning algorithms are used to handle complex driving maneuvers, then motion planning optimality and usability are improved, but computational complexity increases making real-time capability difficult to achieve
Solution Approach 1:
The motion planning problem is segmented into discrete state-space grids, where the continuous state space is divided into manageable discrete cells. This segmentation allows complex maneuvers to be broken down into sequences of discrete state transitions, reducing computational complexity while maintaining planning optimality through systematic exploration of the grid space.
Solution Approach 2:
Motion candidates are pre-calculated and stored in a database before actual motion planning execution. These pre-computed candidates include pre-calculated action discretizations and heuristic rules that are prepared in advance, enabling the system to quickly retrieve and evaluate promising motions without performing complex computations in real-time, thus achieving both optimality and real-time capability.
2Reliability
If comprehensive motion planning algorithms are used to handle complex driving maneuvers, then motion planning optimality and usability are improved, but computing time increases
Solution Approach 1:
Motion candidates and their evaluations are pre-calculated and stored in a database before actual motion planning execution. This preliminary action includes pre-computing action discretizations, evaluating motion candidates against cost functions, and storing heuristic rules. During real-time operation, the system retrieves pre-evaluated candidates rather than computing them on-the-fly, dramatically reducing computing time while maintaining optimality through systematic candidate selection.
Solution Approach 2:
The system evaluates only a selected subset of motion candidates rather than exhaustively computing all possible motions. By using pre-calculated action discretization and heuristic rules to identify promising candidates, the system performs partial evaluation on the most relevant subset, reducing computing time while still achieving optimal or near-optimal motion planning results.
3Loss of time
If action and state gridding with pre-calculated action discretization is used, then computing time is reduced enabling real-time capability, but motion planning precision may be compromised
Solution Approach 1:
The state space is gridded with varying resolution, where finer grids are applied in regions requiring higher precision (such as near obstacles or during complex maneuvers like parking) and coarser grids are used in open spaces. This local quality approach maintains motion planning precision in critical areas while reducing overall computational complexity, enabling real-time capability without sacrificing necessary accuracy.
Solution Approach 2:
The continuous state space is represented by discrete grid copies that approximate the continuous domain. Pre-calculated action discretization creates a discrete representation of continuous actions, and this discrete model is used to evaluate motion candidates. The discrete grid copies maintain sufficient fidelity to the continuous space to achieve accurate motion planning while enabling efficient computation through discretization.
4Loss of time
If heuristics are used to prioritize promising motion candidates, then computing time is reduced, but thoroughness of motion evaluation may be compromised
Solution Approach 1:
Heuristic rules provide feedback mechanisms that guide the selection and evaluation of motion candidates. The heuristics evaluate intermediate states and provide feedback on promising directions, allowing the system to prioritize candidates that are more likely to lead to optimal solutions. This feedback-driven approach maintains thoroughness by systematically evaluating candidates based on multiple criteria including cost functions, constraints, and heuristic scores, ensuring reliable motion evaluation while reducing computing time through intelligent prioritization.
Data Source
AI summary
A method, an apparatus, and a computer-readable storage medium with instructions for motion planning for a transportation vehicle wherein motion candidates are determined based on an action and state gridding. This involves applying at least one measure for optimizing the motion candidates. Final states of motions of the transportation vehicle corresponding to the movement candidates are determined in a discretized state space and a motion for the transportation vehicle is selected.