Formation Control And Experiments Research On Multi-robot Systems

Formation control is a typical and universal multi-robot coordination problem, whichis the foundation of the robot coordination problem and a huge boost and promote for theresearch of the multi-robot systems. It is often desirable that the agents have their ownposition information. In the situation of we cannot obtain the absolute position of agents,we need to propose a formation algorithm which does not need the absolute position ofeach agent and can achieve the efect of formation.In this dissertation, we study the formation control algorithm by using only the relativeposition of agents. The main contents of this dissertation compose as follows:Firstly, this thesis introduces the knowledge of graph which related to the multi-robotsformation control and the modeling method for multi-robots.Secondly, we propose a formation control based on the position estimation which iscombined with the basic knowledge of formation control. Under the proposed strategy,agents can realize the convergence to the desired formation confguration if and only ifthe interaction graph for the agents has a spanning tree. We use the method of positionestimation in one-order, second-order and diferential wheel systems, proposed the diferentcontrol input for them, and has carried the simulation and proof, respectively. For one-order and second-order system, adding the additional input, it can achieve the mutualcollision avoidance efectively.Then, the multi-robot experimental platform builded by our laboratory is presented.Described in detail the composition of the E-PUCK intelligent vehicle, the WI-FI commu-nication system and positioning system based on visual positioning and track estimationwhich developed independent by our laboratory on the basis of E-PUCK.Finally, combined with the formation control algorithm based on position estimationand the multi-agents platform, we conduct formation control experiments on the platformand show the formation experiment results and the data analysis of formation control.