Distributed spot welding task allocation and sequential planning for multi-station multi-robot coordinate assembly processes

Many industrial robots are equipped in the multi-station autobody assembly line to complete the spot welding tasks collaboratively. The task allocation and the sequential planning of the welding spots (WSs) are two key sub-problems and significantly influence the efficiency of the multi-station multi-robot (MSMR) coordination process. However, these two sub-problems are highly coupled and have complex engineering constraints, which makes them hard to be jointly optimized. Traditional methods often optimize the MSMR coordination process through hierarchical optimization, which does not fully consider the coupling effects among constraints and is easy to be trapped in the local optima. In this work, an integrated MSMR task allocation and sequential planning framework is proposed to fully consider the complex engineering constraints (e.g., robot accessibility, collisions, and the cycle time at the station) to model the coordination process. An enhanced biased random key genetic algorithm (BRKGA) is proposed to optimize the proposed framework by explicitly considering the engineering constraints and tackling the local optimality caused by the coupling effects between two sub-problems, in which double-crossover(), double-mutation(), and elite re-optimization() are designed to simultaneously ensure the adjacent robots are assigned distinct sets of welding spots and reduce the time of each robot in completing the welding task. In order to evaluate the effectiveness of the proposed method, the spot welding of the autobody is used as the case study. Compared with the two benchmark methods, the line balance efficiency is improved by 21.087% and 7.803%, respectively.