Design Of Multi-Moble-Robot Cooperation Control Platform And Research On Local Control Algorithm

The consensus of multi-agent systems is widely studied in control community, motivated from their broad applications in science and engineering. And one of the essential problems is how to design a proper test platform used to validate the effectiveness of the consensus algorithm. Moreover, in order to avoid the dangerous states or get to the special states, the system needs to modify its state. However, it is so complex to control every robot for the multi-robot system with a large-scale topology. Therefore, it is significant to change collective behaviors of all agents in the multi-agent system just with local controllers while keeping the system asymptotic state agreement.The key problem of locating all of the agents is solved by processing the images captured from the global camera. And the system asymptotic state agreement switching is achieved by modifying control signals of parts of the agents. The main contributions of this thesis are as follows:1. A multi-robot cooperation control test platform used indoor is designed based on e-puck robot. The computer obtains the state of all the agents by processing the real-time images captured from the global camera. Then control signal outputted by the algorithm model or location information will be broadcast to all the agents in the wireless way. Moreover, the experiments of the centralize control, distributed control and local control experiments can be implemented on this platform.2. A local control algorithm for system state agreement is proposed. The relationship between the strategy and stability of the system has been analyzed, and a necessary and sufficient condition as well as a practical strategy used to select out the better local control strategy are given. The local control can change collective behaviors of all agents in the multi-agent system, and then all of the agents get to the given state, without independently controlling every agent. 3. Based on the above, experiments on the platform are given to validate the effectiveness of the local control algorithm. And these cases are about system with balance graph as well as unbalance graph. Finally, these results show that the local control algorithm is effective and this platform work well.A multi-mobile-robot test platform is designed and the local control algorithm for the consistent balance state switching is proposed. Finally, the algorithm is implemented on the platform, and the experiment results show the effectiveness of the algorithm and the platform works well.