Robot belt grinding trajectory optimization based on GLS-PSO

To automatically generate the reference trajectory of control parameters in robot belt grinding system, this paper presents a Genetic and Local Search-based Particle Swarm Optimization Algorithm to optimize two main parameters, contact force and feed rate. The proposed approach takes advantage of Local Search Technology to accelerate the learning and searching process, which is expected to improve the quality of particles as well; Meanwhile, Genetic crossover between individuals is used to combine good genes to produce better offspring. The experimental results show that the GLS-PSO is superior to LS-PSO, G-PSO and S-SPO in terms of both algorithm performance and optimized effects. In addition, the proposed GLS-PSO algorithm meets the requirements of industrial control in robotic belt grinding, which demonstrates the feasibility of this method.