Multi-robot Formation Control Based On Optimized Behavioral Parameters By Improved Genetic Algorithm

Formation control is a very important problem in multi-robot research area. Multi-robot formation can effective finish the more dangerous and complex task and has a wide application range. Therefore, the research of multi-robot formation has an important theoretical and practical significance. The paper uses behavioral and pilot methods to study multi-robot formation and genetic algorithm to optimize robot behavior of weighting parameters.Genetic algorithm mainly has two problems, the first one is the "premature" phenomenon in the early search period; the other is trapping into local optimal solution in later. To solve the problem, this paper presents a multi parameters adaptive genetic algorithm, the algorithm is to make the genetic algorithm’s crossover probability, mutation probability and the number of mutation gene change automatic with the fitness value of each different individual in the generation of species. Multi parameters adaptive genetic algorithm can maintain the diversity of the species and ensure the convergence of genetic algorithm. The test results show that the improved genetic algorithm has better search optimization ability.The paper uses the behavioral and pilot methods approach to research multi-robot formation control. Firstly, the paper sets up the formation’s geometric center as a virtual leader, who can lead the formation robots to complete the multi-robot formation; Next for each robot, the paper design several basic robot behaviors and use motor-schema method to synthesis the basic behaviors; Finally, the paper uses the improved multi parameters adaptive genetic algorithm to optimize robot behavior parameters. On MATLAB software platform, we establish a simulation environment for multi-robot formation control simulation. Through compare the simulation result can see the feasibility of improved genetic algorithm.