Multi-Manipulator Motion Planning With Dynamic Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for collision-free motion planning in multi-armed robots are inefficient, particularly in dynamically changing environments, as they often rely on probabilistic methods that do not generate deterministic trajectories and can be computationally unpredictable, or local optimization methods that struggle in complex environments.
Innovation Solution
The method employs independent dynamic maps for each manipulator, allowing for deterministic and fast replanning by precalculating mappings of workspace elements to nodes and edges in the configuration space, enabling rapid collision checks and flexible extension to multi-arm systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If probabilistic planning methods are used for multi-armed robots, then computation time can be limited, but the methods do not generate deterministic trajectories and computation time is unpredictable
Solution Approach 1:
The patent pre-calculates and stores mappings from workspace elements to configuration space nodes and edges in dynamic maps before runtime. This preliminary action enables rapid query-based pathfinding during execution, achieving both fast computation and deterministic trajectories by eliminating the need for probabilistic sampling during runtime.
Solution Approach 2:
The patent replaces probabilistic planning algorithms with a deterministic mapping-based approach. By substituting the probabilistic mechanism with a pre-computed spatial mapping structure, the system achieves predictable computation time and deterministic outcomes while maintaining adaptability to dynamic environments.
2Productivity
If local optimization methods are used for motion planning, then real-time planning is enabled, but solutions cannot be found in complex environments
Solution Approach 1:
The patent performs preliminary computation to build comprehensive dynamic maps that capture the full configuration space structure before runtime. This pre-processing enables real-time queries to explore complex environments systematically rather than relying on local optimization, achieving both speed and completeness in complex settings.
Solution Approach 2:
The patent transforms the motion planning problem from direct configuration space search to a workspace-to-configuration-space mapping query problem. By introducing this dimensional transformation through pre-computed mappings, the system achieves efficient real-time performance while maintaining the ability to handle complex environments through the structured mapping relationship.
3Loss of time
If dynamic maps are pre-calculated for each manipulator, then rapid deterministic replanning is enabled, but device complexity increases
Solution Approach 1:
The patent creates simplified copies of the configuration space structure in the form of dynamic maps with pre-computed mappings from workspace to configuration space. These copies enable rapid querying without requiring the full complexity of the original configuration space representation to be maintained in memory during runtime, achieving speed while managing complexity through selective replication.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method (100) for collision-free movement planning for a first manipulator (12) in a first workspace (16) and for a second manipulator (14) in a second workspace (18), wherein the first and second workspaces (16, 18) overlap at least in part. The method comprises reading in a first dynamic map (22) for a first configuration space of the first manipulator (12), wherein the first dynamic map (22) comprises a first search graph and a first mapping between the first workspace (16) and the first search graph, reading in a second dynamic map (28) for a second configuration space of the second manipulator (14), wherein the second dynamic map (28) comprises a second search graph and a second mapping between the second workspace (18) and the second search graph, and coordinating the movement of the first manipulator (12) and the second manipulator (14) based on the first dynamic map (22) and the second dynamic map (28).