Distributed Optimization And Control Algorithm Of Flocking For Swarm Robots System

For significant advantages on the robustness, adaptability, extendibility and economy of the swarm robots system, it has been an active research in the robot area in resent years. In this thesis, the main research work is the adaptability to the complex environment of swarm robots system, the controllability of swarm velocity and the optimization of system performance, which is based on the flocking control of swarm robots system. This work is supported by the National Nature Science Fund Project and Graduate Innovative Project. The main work in this paper is as follows:Firstly, the distributed control strategy is designed to make the flocking behavior of swarm robots system adapt to nonlinear environment. And the nonlinear environment is described by a given potential function which is assumed to have finite bounded slopes. And based on the local information interaction distributed control algorithm is presented by using the idea of virtual force and nearest neighborhood law, which can realize the swarm flocking behavior in the nonlinear environment, and the finish time of flocking can be estimated. Simulation results are included to verify the distributed controller, and also show the adaptability and extendibility of swarm system.Secondly, the controller is designed for swarm robot system in the environment with unknown obstacles and the stability of system is proved. The traditional artificial potential field method is improved to achieve obstacle avoidance based on the number advantage of the swarm. After perceiving obstacles the swarm can achieve safe avoidance with the obstacles and then finish flocking task by using the heading effect of virtual leader in swarm. To validate the proposed algorithm, the theory analysis and experimental simulation are performed. And the results show that the algorithm can make swarm avoid obstacles effectively, and overcome the defect that traditional artificial potential field method easy to deadlock.Then, the velocity controllability problem of swarm flocking is solved without breaking the local interaction rules of swarm robots system. Controllable agents (one or more) are introduced to the swarm based on the idea of "soft control", and the velocity controllability of the swarm is realized through controlling the controllable agents. For the swarm autonomous system with constant expected speed and the swarm non-autonomous system with time-varying expected velocity, LaSalle invariance principle and Barbalat Lemma is adopted in analysis of system stability, so that we proof speed vector of all robots can converge to the desired value, the swarm will form a stable structure, and the more controllable agents are introduced, the faster the swarm convergence. So the results show that the designed control strategy is effective.Finally, the synthesis performance optimization problem of consistency control for groups of autonomous robots system is analyzed and studied. Make the velocity error of swarm, the structure error of swarm and the traffic as optimization goal, and treat the minimum convergence time of swarm velocity as constraint, particle swarm optimization algorithm is used to obtain the best communication frequency and communication radius of swarm, in order to improve the accuracy of controlled state, the speed of swarm convergence, and the economy of communication costs.In summary, a series of theoretical studies about the flocking behavior of swarm robots system are conducted in this thesis. The main purpose of this work is to realize the robustness, adaptability and extendibility of swarm flocking system. And then, Simulation experiments are performed for the purpose of related verification and analysis.