Research On Formation And Consistency Control Of Multi-robots Based On Improved Particle Swarm Optimization

Multi-robot system has attracted considerable attention and many good results has been obtained in recent years. The research contents in this field mainly include task allocation, path planning and optimization, map building, formation control, target search and seizure, learning strategies, and so on. Particle swarm optimization has been employed in the control and collaboration optimization of multi-robot system. Particle swarm optimization is a kind of intelligent algorithm, which does not need to establish mathematical model and works through particles mutual learning as well as the usage of historical information. Based on improved particle swarm optimization strategy and improved virtual structure method, the goal of this thesis mainly focuses upon the study on the consistency motion control, formation control. The main work of this thesis is given as follows.Firstly, comparison analyses among particle swarm optimization, ant colony algorithm and neural network are done based on the interpretation of basic principle of particle swarm optimization and working mechanism of its parameters. The status about particle swarm algorithms applied in the field of multi-robot research is reviewed, and the outlook over the coordination control of multiple robots working on community perception networks is proposed.Secondly, considering the uncertain unknown environment information about shapes, sizes and positions of obstacles, a modified particle swarm optimization scheme with the advantage of good local optimization virtue is proposed to achievement the destination consensus of multi-robot systems by employing the three fitness components via destination information, ambient environment information, robot motion information. Finally, the simulation experiments are demonstrated the effectiveness and feasibility of the proposed strategy.Thirdly, an improved virtual structure method based formation motion control scheme is put forward for the difficulty of formation expression and realization problem of multi-robot formation task. By introducing virtual navigation robot into building the desired virtual formation, an asynchronous matching strategy among real robots and virtual robots is proposed. Based on this, an improved particle swarm optimization formation scheme is given with the help of a relatively optimal robot in obstacle avoidance selection environment.Finally, the community perception networks are built through the random deployment of static intelligent perception nodes according to the situation of wide working area and more robots to formation control. Based on an improved particle swarm optimization idea, a novel formation control is provided in the community perception network environment. By designing the community network sub-formation generation method and obstacle avoidance scheme for sub-formations with the employment of artificial potential field method, the final formation integration strategy on community perception networks is proposed.