Multi-level optimization approach for multi-robot manufacturing systems

Abstract:
This paper presents a multi-level optimization approach for multi-robot system in the context of Software Defined Manufacturing (SDM). While industrial robots offer larger workspace, versatility, and lower cost, they generally lack in rigidity and thus accuracy in comparison to machine tools. We propose a solution to compensating these drawbacks using a multi-level optimization which includes an intelligent task and action planning for multiple robots, a trajectory generation considering robot dynamics, coupling of robots, and stiffness constraints, as well as a modular simulative validation. Finally, we present an experimental setup with two coupled robots in a bending task. The results show this approach allows robots to apply significantly higher forces compared to single robot setup, and dual robot benchmark without stiffness optimization.