Gpu-accelerated planner improves multi-arm robot coordination efficiency
cuArms: GPU-Accelerated Conflict-Based Planning for Multi-Arm Manipulation
RoboticsArtificial IntelligenceMultiagent Systems
Summary
Controlling several robot arms to work together can make tasks faster and more flexible, but it's usually slow and complicated. The authors present cuArms, a tool that speeds up this process by using graphics cards to quickly plan and avoid conflicts between arms. Their approach can quickly create coordinated robot motions without needing long training times. This means robots can better share workspaces and handle tasks more effectively.
What this means in practice
- •For industrial robot integrators: Generate coordinated motion plans for multiple robot arms in shared workspaces with faster initial solutions and ongoing improvements.$Commercial implications: Enables the sale of efficient multi-arm coordination software for manufacturing and automation by speeding up planning without extensive training.
- •For warehouse automation teams: Improve scheduling and coordination of robot arms handling diverse manipulation tasks to increase throughput and reduce conflicts.
Authors
Philip Huang, Fabio Ramos, Jiaoyang Li, Caelan Garrett, Lars Johannsmeier
Abstract
Multi-arm manipulation has the potential to dramatically improve the throughput, efficiency, and dexterity of many manipulation tasks. Recent work coordinates multiple robot arms in a shared workcell by allocating and scheduling tasks for greater parallelism, jointly planning multi-arm motions, or directly learning multi-arm policies. So far, these approaches are either too slow, rely on domain-specific assumptions, or scale poorly with the number of robots and task horizon. Thus, we introduce cuArms, a conflict-based planner co-designed with GPU parallelism that efficiently decomposes the search space using reusable multi-modal roadmaps for individual robots and resolves conflicts and deadlocks through a lazy search. To jointly optimize task assignment, schedules, motions, and manipulation skills, we integrate a high-level scheduler that proposes candidate skeletons for cuArms to ground into coordinated manipulation plans. Experimental results show that our method establishes a new state of the art, with real-time initial solutions and anytime improvements in both motion quality and scheduling, without days of domain-specific training required by learning approaches. Videos of both simulated and real-world industrial tasks are available on https://cuarms.github.io/.