Multi-Objective Optimization of Spot-Welding Robot Trajectories Using an Improved NSGA-II Algorithm

To address the inherent drawbacks of conventional spot-welding path-planning methods, which cannot simultaneously consider welding efficiency, energy consumption and trajectory motion stability, this paper proposes an intelligent path-planning approach for spot-welding robots based on the improved NSGA-II multi-objective optimization algorithm. First, the NSGA-II algorithm is optimized with improved targeting to the spot-welding process. An integer-permutation encoding scheme together with a random initialization strategy is adopted to guarantee the feasibility of solutions for welding-point sequences. Meanwhile, the original genetic operators are replaced. Ordered crossover and Swap mutation are employed to enhance the algorithm’s adaptability to combinatorial optimization problems. Second, focusing on the core evaluation indicators for robot paths, including path length, energy consumption and motion stability, this paper constructs a quantitative characterization model for each indicator and integrates these models into the multi-objective optimization framework. On this basis, path-simulation experiments are carried out on the JAKA-A5 robotic platform. The experimental results show that the proposed method can realize the collaborative optimization of the above-mentioned objectives. Compared with the unoptimized random sequence, the path length is reduced by 36.67%, the energy consumption is decreased by 75.89%, and the motion stability is improved by 11.72%.