The Formation Control For Networked Multi-agent Systems And Experimental Validation

Networked multi-agent systems are systems where the information exchange among agents is accomplished via networks.Compared with traditional control systems,multiagent systems render advantages such as parallel computation,fault tolerance,robustness to disturbances,and etc.Due to its potential application in space exploration,oceanography and etc,networked multi-agent systems have become an extremely active topic in the control community.Formation control is defined as the agents keep a particular geometry configuration in the movement subjected to some constraints.The formation control problem has potential value in aerospace exploration and industry,therefore this problem has important value in both theoretical research and engineering.In this paper,the author designs a complete multi-layer formation control strategy which covers the control allocation problem for a single agent and a coordination controller for multiple agents.The author carried out tasks in the four following aspects:First,the information exchange among agents is accomplished via networks which bring communication time delays and data dropouts to the overall system,therefore,the performance of the overall system will be affected.In this thesis,a novel practice-related networked predictive control method is proposed,and the control method is applied to networked linear multi-agent systems.Compared with the existing networked predictive control methods,the proposed method renders advantages such as high computational efficiency,being easy to be implemented,disturbance rejection and etc.The formation performance and the stability of the overall system are analyzed and the conditions to achieve the control goals are derived.At last,experiments based on air-bearing testbeds are implemented to test the performance of the proposed algorithm and the correctness of the theoretical analysis.Second,for a single agent,the parameter uncertainty of the thrusting system will drive the practical control effects(e.g.,forces and torques)deviated from the reference signals thus cause deterioration of the control performance.Two adaptive control allocation schemes are proposed for nonlinear plants with parameter uncertainty.The stability analysis is given.These two schemes improve the control performance of the system without measuring control effects(forces and torques).Compared with traditional methods,the method based on model reference adaptive control relaxes the conditions to apply the algorithm.The author designs experiments based on air-bearing testbeds to test the performance of the proposed schemes and the correctness of the theoretical analysis.Third,controlled plants in practice are generally nonlinear systems,however,the proposed networked predictive control strategy is designed for linear systems.A novel networked predictive control scheme is designed for the coordinated path-following control problem of nonlinear autonomous vehicles.By applying cascade system theory and passivity theory,the stability analysis is completed.Simulations and experiments are implemented to test the performance of the proposed algorithm and the correctness of the theoretical analysis.At last,the control strategies for networked multi-agent systems are mostly validated by simulations,very few algorithms are verified by experiments,which makes the theoretical results hard to apply in practice.To validate the effectiveness of the proposed control schemes in this paper,the author designs a wireless light-weighted networked control platform based on android smartphones which shortens the developing period,meanwhile provides functionalities such as supervisory and monitoring of the controlled plants on-line.The experimental results validate the practical value of the proposed control schemes.In this thesis,networked predictive control schemes are proposed for both linear multi-agent systems and nonlinear multi-agent systems,the stability and the control performance of the overall systems are analyzed.In the experiments based on air bearing testbeds,the author designs applicable wireless controllers to validate the effectiveness of the proposed control scheme and the correctness of the theoretical analysis.