Multi-Objective Trajectory Optimization of Collaborative Robots Based on improved NSGA-III

In industrial production, collaborative robots require joint trajectories to meet multiple optimization objectives such as time, energy consumption, smoothness, and safety, to ensure the collaboration with humans. Therefore, this paper addresses the multi-objective trajectory optimization problem for collaborative robots, focusing on achieving the optimal time, energy efficiency, and smoothness. A method based on the NSGA-III is proposed to improve the issues of slow algorithm convergence and susceptibility to local optimal solutions. Firstly, to meet the requirement of smooth continuity of robot joint trajectories, seventh-order B-spline curves are used for planning to ensure the continuity and stability of robot motion. Secondly, considering the constraints of joint trajectories and robot kine-matics, emphasis is placed on the work efficiency, lifespan, energy consumption, and motion impact of the robot, proposing multiple optimization objectives including minimizing time, optimizing energy, and minimizing the motion impact. Finally, based on the characteristics of multi-objectives and the definition of the Pareto optimal solution set, the improved NSGA-III is utilized to solve the multi-objective trajectory optimization model. We further avoid the problem of the original algorithm prematurely falling into the local optima. Experimental results show that the optimized robot joint trajectories have shorter running time, lower energy consumption, and reduced motion impact. The continuity and smoothness of joint velocities, accelerations, and impact trajectories provide strong assurance for the stability and lifespan of robot motion.