Motion Synchronization Control And Cooperative Task Planning For Multiple Mobile Robots Network

As two of the most important issues in the researches of multiple mobile robotics,motion synchronization and task cooperation are the foundation of the accomplishment of several highlevel applications and complex tasks such as environment exploration,rescue,cooperative target tracking and material transportation.Motion synchronization requires that the speed and orientation of each robot maintain synchronization with those of its neighbors during the whole motion process,which guarantees that proper formation shapes are maintained in the robot group level,to decrease energy consumptions as well as improve the completion efficiency and robustness of coordination tasks.In task cooperation,the complicated task is firstly divided into several sub tasks and then be allocated to proper robots according to their different capabilities and current states.Each robot resolves its motion planning according to the assigned tasks and maintains coordination with its neighbours,eventually the complex task can be accomplished by the robot group through cooperation.This thesis investigates the motion synchronization control and cooperative task planning problems of multiple mobile robot network and presents a series approaches on the system modelling,task planning,control design and performance analysis,mainly including the following three aspects:1.For the synchronous formation control problem of multiple robots under the network communication environment,a fully distributed control approach as well as the corresponding stability analysis are presented.Firstly,based on the establishment of the synchronous formation model(including the time-varying formation model,robot desired trajectory description and motion synchronization constraints),the basic synchronous formation control algorithm is presented.Then,in order to solve the problems of data sampling and communication delay in information exchange channels,the synchronous formation control system under sampled-data and communication delays is formulated and the corresponding distributed controller design is proposed.Based on the Lyapunov-Krasovskii approach,a sufficient condition is further presented which guarantees the exponential convergence of the proposed synchronous formation control system.Furthermore,in order to solve the practical problems that the robot ability has limitations and model parameters can not be estimated accurately in real applications,the proposed distributed synchronous formation control algorithm under sampled-data and communication delays is further extended to solve the formation control problems without coupling velocity feedbacks and under parameter uncertainties in the dynamic model.Finally,simulation and experimental results validate the effectiveness and the robustness of the proposed synchronous formation control approach.2.In order to resolve the synchronization decline problem caused by robot failures during the synchronized motion process,a gradient-based recursive self-repairing approach is presented which can repair the network topology at the same time,restore the motion synchronism of the robot formation.The proposed approach utilizes the relay switched topology control based recursive self-healing framework.More specially,by introducing the gradient generation and diffusion mechanism,the proposed approach guarantees that each robot in the formation has the ability of distributed distance estimations,then the optimal repairing path can be achieved in the self-repairing process under the fully distributed control framework.Simulation results validate that the proposed approach can repair the network topology as well as restore the motion synchronization performance of the robot formation.Compared with existing algorithms,the proposed approach has less repair robots and shorter repair path.3.In order to resolve the planning problem of the complex and cooperative tasks of multiple mobile robots,by taking the multi-robot transportation task in industrial assembly and warehousing logistics applications as a typical example,an incidental delivery oriented cooperative task planning approach is presented.Firstly,from the viewpoint of robotics,the mathematical model of the cooperative transportation tasks is presented and the cooperative task planning problem is formulated in a standard optimization framework.More specially,the incidentally task concept as well as its mathematical model are firstly presented in this thesis,and based on which the incidental delivery based single robot route planning algorithm is presented,which guarantees the optimal route planning results in the single robot level.Secondly,according to the different requirements of the cooperative transportation problems,a simulated annealing based hybrid high-level task allocation algorithm and an auction based distributed high-level task allocation algorithm are designed,and then incorporating the proposed single robot route planning algorithm,the integrated task planning approaches of the multi-robot cooperative transportation problem are presented.Furthermore,in order to solve the dynamic task planning problem caused by robot failures,task failures and online task publishing,a fully distributed online task planning algorithm is presented with its optimization performance analyses.Simulation and experimental results validate the effectiveness,real-time performance and scalability of the proposed approach,as well as its application prospects in practical applications.