The Research Of Collision Avoidance And Task Cooperation For Multi-Robot

Multi-robot system has shorter designing period, higher stability and efficiency, better robust to environment than single robot with high intelligence, which makes it to be an important part of intelligent robot research. However, if it does not have an availably cooperative algorithm among robots to coordinate themselves, the efficiency of performing the task will be highly diminished, even interruption or collision between robots may happen. Therefore, the research on multi-robot cooperation and coordination will not only improve the intelligent performance of robot, but also accelerate the practical process of multi-robot task system.The architecture of multi-robot system, motion planning, task cooperation and communications etc, were deeply researched in this thesis. Every function module was designed, and a set of cooperation and coordination methods for the unknown dynamic environment were brought forward.Firstly, the whole architecture of multi-robot system and some related coordination strategies was introduced generally. Secondly, a kind of mixed coordination control strategy based on the improved potential field method and predictive avoidance rule was brought forward. For the problem of dead lock in process, valid solutions were added. Thirdly, the mode of market mechanism and negotiation was used to accomplish task of allocation and cooperation in the design of multi-robot's cooperating model. Some functions such as roles state changing of robot, task organization and allocation, the negotiation and cooperation among robots were realized by the model.Finally, the simulative platform of multi-robot system was established based on Visual C++ language. The versatility of this platform was verified by the simulations under unknown complicated dynamic environments, and the validity of multi-robot cooperation algorithm was also proved.