Formation And Dynamic Obstacle Avoidance Control Based On Community Networks

With the progress of multi-robot research, the coordination control in unknown environment has become a hot topic. Although the study of collaborative control among robots in unknown environments has achieved mang good results, but, multi-robot system is still facing many chanllenges such as limited sensing range, intermittent communication between multiple robots and the interaction with the external environment real-time is not very well.Fuzzy control is a practical method to deal with multi-robot system coordination problem in unknown environment. By basing on existing expertise, through the establishment of different input and output parameters of membership functions and fuzzy rules corresponding to achieve better motion control of multi-robot systems in different environments. By arranging static perception nodes in an unknown environment, community perception network can be built through wireless communication. The main problems about multi-robot obstacle avoidance, path-following and formation in unknown environment were foucused upon in this thesis, by employing fuzzy control method, which has good value and important significance for robot system. The main work in this thesis is as follows.Firstly, for the problem of path following in a class of unknown environment which include static obstacles, a hierarchical fuzzy path following control scheme based on different fuzzy grain size was presented which based on different fuzzy grain.Secondly, for the dynamic obstacle avoidance problem in a class of unknown environment, according to the speed and direction change information of dynamic obstacles, the fuzzy rules of the delta speed and deflection angle of robot was established, a second-order fuzzy control program is proposed based on fuzzy method.Thirdly, for the multi-robot formation control problem at an unknown uncertain environment, due to the difficulty of effectively maintainance of the multi-robot formations, different fuzzy formation keeping rules and different formation changing rules were designed to achieve the effective control of three kinds of multi-robot formation tasks.Fourthly, considering the limited perection range of sensors, community perception networks was constructed, by arranging the static sensor nodes in the working environment. Leader robot of path planning and multi-robot formation fuzzy control scheme were proposed by employing grid method and the visible image method.