Research On The Coordination And Groups Gaming Of Multi-robot System Under Competitive Environments

In recent years, the flocking motion control of multi-robot has attracted more attention ofresearchers from domestic and abroad. It will have important realistic significance to explorethe coordination control strategy of multi-robot between the individuals, especially byintroducing the competitive game theory into the study of the coordination control or flockingmotion planning of multi-robot. Research on how to organize group composed by multiplerobots and how to realize the coordination and cooperation of the robots in the group is ofimportant theoretical and realistic significance. This paper mainly studies coordinated controlproblem of game group of multiple mobile robot in against environment. Firstly, acoordinated control problem of multi-robot in against environment and a collaboratedformation control problem of multi-robot group system are studied from game behavior ofpursuit-escape among two robot groups, then further study three dimensional (3D) spacepath planning problem of single mobile robot. The main results and contributions of thisdissertation are as follows:Firstly, the related technologies of mobile robot are introduced and the research methodof coordinated control of multi-robot group and the research situation of formation controland path planning technology are summarized, laying the root for follow-up research.Secondly, aiming to the problem of the trajectory unknown for the target robot whichescapes on its own initiative, a multi-robot collaborated hunting algorithm is proposed basedon dynamic prediction of the target in dynamic environment. The states of the target robot areobtained to fit its trajectory and the consensus based on consensus-based Kalman Filteringis further used to dynamically predict the position of the target. Then the appropriateencirclement points are distributed for the pursuit robots by the particle swarm optimizationalgorithm.The pursuit robots constantly contract the encirclement for the target robot so thatthey finish the task.Thirdly, the federation search of multi-robot group through formation has realsignificance to improve the efficiency of cooperation in pursuit-escape game. A formationformed and maintained problem of multi-robot group is researched. Four-wheel cardan drivenmobile robot in the laboratory are the studying subjects. The single mobile robot kinematicsmodel is established, based on graph theory, potential function related to multi-robot systemand structure is established and a distributed control algorithm for arbitrary shape ofmulti-robot is discussed. Then a kind of distributed control law that can effectively control theformation of robot is designed based on lyapunov stability theory and the sufficient conditions that the formed formation of multi-robot system is stable and the control problem has solutionis given.Finally, the hierarchical path planning behavior of the individual robot is considered onthe basis of the research on the game and coordination of multi-robot system and the pathplanning space is extended to3-D environmental space. A hybrid variation particle swarmoptimization (HPSO) algorithm is proposed to solve global path planning problem for mobilerobots in3-D limited space. Firstly, it reduces the dimension of the3-D space and furtherdivides the ‘feasible region’and ‘search boundary’, which narrows the search area of theshortest path solution. The particle swarm are shrunk towards the center axis from start pointto target point of planning in the process of optimization to press closer to the border ofobstacles by using the directional variation and random mutation to contract solution space,where the planning paths are relatively concentrated and are basically distributed in the centerline or so, so that they will be closer to the optimal solution. The simulation results prove thesuperiority of the presented path planning method in3-D environmental space for the robots,so that it lays the foundation for the applied research of path palnning of multi-robot group in3-D space.