Abstract: In view of problems such as difficulty in path planning of logistics handling robots in complex environments, a trajectory optimization control method for handling robots based on membrane computing and model prediction control was proposed. Firstly, an improved robot dynamic window algorithm based on membrane computing particle swarm optimization was proposed to address the problems of low efficiency in trajectory planning caused by the traditional dynamic window algorithm using uniform and equal division sampling in the speed sampling space. With the help of the randomness of particle swarm optimization and the distributed parallel computing capability of membrane computing, traditional dynamic window algorithms were optimized and iteratively optimized to obtain the optimal path. Secondly, in response to the nonlinear characteristics of the robot system model, a model predictive control method was designed to complete trajectory tracking, by building prediction models, setting objective functions, and eliminating cumulative errors through integration. The experimental results showed that the improved algorithm reduced the average path length, time, and number of steps by 15.07 %, 6.72 %, and 7.68 % in both sparse and complex obstacle scenarios,the proposed model predictive control method has certain advantages in tracking accuracy and system robustness.

Key words: handling robots, membrane calculation, Particle Swarm Optimization (PSO) algorithm, Dynamic Window Algorithm (DWA), model prediction control