The Research On Formation Motion Of Multi-robot System

This paper considered a problem about formation of multi-robot system. The problem includes two tasks. The one is about formation, the other is about obstacle avoidance and path planning. The related algorithm research, design, theoretical proof and simulation have been carried out respectively.To the first task, the research starts from one leader-follower fonnation, sums up the model of the car-like robot, and transforms the unicycle robot system model into the tracking system model by introducing virtual robots. The input-output feedback linearization has been used to design the controller and the perturbed system theory has been used to prove its stability. Three propositions put forth afterward have been proved by the strict theoretical derivation. The correlational algorithms have been demonstrated by Matlab. Then the formation structure of the small scale system is described by the graph theory. The researches on relevant theoretical are carried out on large-scale multi-robot formation description and transformation method.To the second task, the drawback of the obstacle avoidance and path planning using the current artificial potential method are analyzed, that is obstacles have not been described properly and to choose the algorithm parameters is very complex. To solve the two problems, the Region Potential Field Switch method have been proposed and by using it, the reasonable description of obstacles in map is given. It is successful to plan the right path under the method.Finally, by analyzing the system architecture from the perspective of global and merging the various behaviors of formation through behavior method, the task about system formation is implemented. All the algorithms the paper proposed are verified by Matlab simulations. It is successful to complete the formation of multi-robot system.