Research On The Control Of Multi-robot Formation

With the development of robot technology, coordination of multi-robot technology has become an important direction and the most basic is the formation of multi-robot. In this paper, the formation methods of multi-robot control were studied.Formation has many areas of application. This paper focuses on the advantages of multi-robot formation, pointing out the need to pay attention to the control group and comparing with the current focus on the formation of several well-known methods of control. In the control, the choice of system architecture is very important. Multi-robot system uses hybrid architecture and the robot individual architecture is based on behavior. To improve the output behavior's pertinence, we have designed a robot based on the behavior of single-layer structure of the integration of information and the individual robot-integration architecture.The formation control combines with the use of hybrid approach: behavior-based. It designs four behaviors such as toward the goal, maintaining the formation, avoiding obstacles and avoiding the other robots. This article introduces the formation conditions of feedback mechanism in designing maintaining formation behavior. To avoid obstacles, artificial potential field with the leader-follower is used firstly. Secondly, we design the method of division access domain. Formation deformation and formation change obstacle avoidance methods are proposed. Those followers access to global information through the blackboard which has leader's issues. At last, the paper designs a multi-robot formation mechanism of fusion acts.Finally, this article unites Visual C + +6.0 and Matlab6.5 programming to prove the validity of proposed methods, making some simulations and detail analysis of the experiments on a single robot motion planning, the formation of multi-robot formation, with the formation of multi-robot control, more effective obstacle avoidance robots. In order to control the formation and the combination of specific tasks, this paper uses finite state machine concept designing round up the goal of multi-robot simulation.