Multi-Agent Path Planning With Group-Aware Conflict Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current multi-agent path planning algorithms face challenges in efficiently planning motion paths for multiple agents in environments with multiple objects and groups, as they often fail to guarantee optimality and completeness, especially when considering interactions and group affiliations.
Innovation Solution
A computer-implemented method using a central unit for coupled path planning, employing a conflict-based search algorithm with a constraint tree and social force-field computation to determine motion paths that avoid conflicts with objects, individuals, and groups, by incorporating interaction costs and force fields to guide agent movement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If decoupled path planning is used for fast parallel computation, then computation speed is improved, but optimality and completeness are not guaranteed
Solution Approach 1:
The patent segments the path planning problem into hierarchical levels: high-level conflict detection and resolution, and low-level individual path optimization. This segmentation allows parallel computation at the low level while maintaining global optimality through coordinated conflict resolution at the high level, thus resolving the contradiction between speed and reliability.
Solution Approach 2:
The patent introduces conflict constraints as intermediary elements that mediate between individual agent path plans. These constraints act as a coordination mechanism that ensures global optimality and completeness while allowing parallel computation of individual paths, thus maintaining both speed and reliability.
2Reliability
If coupled path planning searches the entire state space for all agents, then optimality and completeness are guaranteed, but computation time increases significantly
Solution Approach 1:
The patent segments the state space search into agent-specific subspaces with conflict constraints. Instead of searching the entire joint state space, each agent's path is optimized independently subject to conflict constraints, dramatically reducing computation time while maintaining optimality through the constraint-based coordination mechanism.
Solution Approach 2:
The patent performs partial search by focusing computation on individual agent paths with conflict constraints rather than the complete joint state space. This partial action approach achieves sufficient optimality for practical applications while avoiding the exponential computation time of exhaustive search.
3Device complexity
If traditional path planning considers only individual obstacles, then path computation is simpler, but paths between group members are not avoided
Solution Approach 1:
The patent merges individual obstacle constraints with group conflict constraints into a unified constraint framework. By combining these constraint types, the system maintains relative simplicity while achieving reliable group conflict avoidance, as the merged constraints are processed through the same conflict resolution mechanism.
Solution Approach 2:
The patent creates a universal constraint framework that handles both individual obstacles and group conflicts through the same conflict detection and resolution mechanism. This multi-functional approach manages increased complexity while ensuring reliable group conflict avoidance.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention relates to a computer-implemented method for planning a motion path for multiple agents (2), comprising the steps of: - performing (S4) a conflict-based motion planning for the multiple agents (2), wherein conflict-free motion paths for each of the agents (2) are determined depending on movement costs, - determining (S2) the poses and velocities of one or more individual objects (I) and one or more groups (G) of objects; - calculating (S3) the movement costs depending on interaction costs of each of the agents (2) with the one or more objects (I) and/or the one or more groups (G) of objects.