Research On Multi-robot Path Planning Based On Multi-objective Optimization

With the development of automated logistics system,the mobile robot being taken as a key tool of transportation system,the technologie in all aspects have been rapidly developed.Path planning for multiple mobile robots is a challenging problem in the field of robot navigation.Path planning for multiple mobile robots is a multiple objective optimization problem,the research difficulties of this problem are how to realize the collision prevention among robots,the obstacle avoidance between robot and barrier and how to guarantee multiple objective optimization of path planning be optimized simultaneously and reasonably.In this thesis,the multi-robot system in the automated logistics system is studied.The current multi-robot path planning technology is analyzed,and the overall design of two-layer path planning scheme is completed,on the basis of the first layer global path planning,the second layer path coordination strategy is added.In the problem of global path planning,most algorithms have the disadvantages of slow convergence speed and insufficient convergence,and most of them only consider the goal of a single path length,resulting in the problem that the path is not smooth and the safety performance is not good.In this thesis,a beetle swarm optimization algorithm(BSO)combining the beetle algorithms search(BAS)and the particle swarm optimization(PSO)is proposed,and a weighted sum method is applied to multiple goals to achieve the optimization of multiple goals.By comparing the path planning results of the BSO algorithm,genetic algorithm and adptive inertial weight particle swarm algorithm,it is proved that the algorithm proposed in this thesis accelerates the search speed,and the obtained path has better overall smoothness,shorter path and farther distance from obstacles.In the further study of BSO algorithm,it is found that the multiple targets cannot be balanced well for the path obtained by the weighted optimization method.Therefore,on the basis of BSO algorithm,the Pareto concept is introduced,and the multi-objective beetle swarm algorithm(MOBSO)is proposed,so that the multiple solutions can be obtained by the algorithm.For the problem of insufficient diversity and insufficient convergence to the Pareto front in most of multi-objective algorithms,an external archiving strategy of infeasible solution is introduced into the path planning,combining infeasible path through obstacles with feasible path,the edge search is enhanced to obtain a better solution.The improved algorithm is verified by the simulation of a single robot,more feasible paths can be planned,and the distribution and convergence of paths are better.In the second level path collision avoidance coordination,aiming at the lack of robot safety in the multi-robot path coordination,the priority algorithm is improved and the strategy of restricted area is proposed,so that the robots can avoid the collision more effectively.The simulation results have shown that the improved algorithm is feasible to solve the collision avoidance of multi-robot,and by compared with the standard algorithm,it is verified that the safety performance of the robot can be improved by the improved algorithm.The application of the improved global path planning algorithm and collision avoidance algorithm in the actual robot path planning is realized through the experiments conducted on the built multi-robot experimental platform,and the effectiveness of improved algorithm have been proved.