The application of hybrid orthogonal particle swarm optimization in robotic path planning

Since particle swarm optimization is easily to fall into local optimization in robotic path planning, we proposed a novel hybrid orthogonal particle swarm optimization method to overcome the drawback. In the proposed algorithm, the global search ability was enhanced by introducing orthogonal design operator to the basic PSO algorithm and a special method to initial particles was designed to find the feasible path. Compared with traditional algorithms, the simulation result of our method demonstrated the superiority of the enhanced algorithm.

I. introduction

Path planning is a key technology for mobile robot and it is also a focus problem of robotic research. Particle swarm optimization algorithm as a new intelligent optimization algorithm has been used to solve this problem due to the merit of rapid searching and easier realization. But the results did not approach ideal consequence, because it was easy to trap into local optimum [1],[2]. As orthogonal experimental design method has a good ability of global search, it has been implicated in genetic algorithm to deal with various optimization problem by many researchers recently^[3],[4]. In Xue's [5], an orthogonal initial method was proposed to solve function optimization problems. Based on the research of orthogonal experimental design method, a new hybrid orthogonal particle swarm optimization (HOPSO) was proposed, which hybridized the orthogonal design operator into the process of particle swarm optimization to solve robotic global path planning problem. In this paper, the model for the global path planning problem was built at first, and then the design scheme and specific realization were given. Finally the rationality and validity of the algorithm was analyzed based on the simulation experiments and the results.