Near-Optimal Assembly Task Sequencing and Allocation Method for Multi-Arm Robot System

We propose a novel approach for near-optimal assembly task sequencing and allocation for a multi-arm-robot system; the problem is known as an NP-hard problem. The challenging issues are solving the geometrical constraints between the parts to determine feasible assembly sequences without collision and avoiding collisions between the assembled parts and the assembling robots as well as robots mutual collisions. In this paper, the combinations of assembly tasks and robots are evaluated in assemblability, mutual collision, and assembly cost at every sequencing and allocation step to discard infeasible or higher-cost task allocations to keep the problem space small to solve. The approach is verified by solving feasible assembly task sequences and allocations of multi-arm robot systems for a twenty-six-part EV battery model and simulated on ROS.