Research On Consensus Of Multiple Manipulators With Elastic Joints

Recently,with the rapid development of multi-agent systems,many researchers have begun to pay attention to multi-agent systems.The research of multi-agent systems mainly includes consensus,formation,rendezvous,optimization,etc.The consensus problem has been deeply and extensively studied as a basic research direction in the field of coordination control.The fundamental purpose of consensus is to use local communication to design control algorithms with which agents' states achieve the same final value.Existing results on consensus of multi-agent systems often focus on linear agent dynamics.However,more and more manipulators with elastic joints which are built with light weight and large load ratio are put into practical use,and existing consensus algorithm cannot be applied directly.To solve these problems,this thesis proposes several consensus algorithms for multiple manipulators with elastic joints.The main results can be characterized as follows:Under the condition that the communication topology graph is an undirected graph,the consensus problem of the multiple manipulators with elastic joints is proposed.The dynamics of manipulator with elastic joints can be regarded as a cascade system.We first design a consensus algorithm associated with the angles of links.Then,following the idea of backstepping,the actual control input on rotor is designed.Based on the Lyapunov stability theory,we show the angles of of multiple manipulators with elastic joints achieve consensus asymptotically.Motivated by the fact that relative velocity information and acceleration information are difficult to obtain in practical situations,a consensus algorithm that requires only relative position information under a directed graph is proposed.A linear reference model is designed for each manipulator,and the model reference adaptive consensus method is used to transform consensus problem into two subproblems,namely,the consensus of reference models and trajectory tracking problem of single manipulator.We derive the consensus algorithm by backstepping method,and it is shown that the angles of manipulators achieve consensus asymptotically.Simulation results are demonstrated to verify the effectiveness of proposed algorithms.We further consider the case in the presence of parametric uncertainties under a directed graph,where an adaptive algorithm is proposed.Combining the dynamic surfacecontrol and adaptive control,a tracking control algorithm is designed to avoid excessive differential terms.Finally,we use Lyapunov stability theory to prove the semi-global stability of the system.