Model Free Adaptive Multi-agent Systems Coordinated Control With Applications In Traffic Networks

In this dissertation,based on the model free adaptive control theory,novel distributed data-driven controllers are designed to address three types of fundamental issues in multi-agent systems.Then,considering traffic flow repetitiveness,the model free adaptive iterative learning control method is applied to the multi-region urban traffic networks.The main contents and contributions of the dissertation are summarized as follows:1.A novel distributed model free adaptive consensus controller is proposed to deal with the consensus problem for a class of unknown nonaffine nonlinear heterogeneous multi-agent systems.First,the follower agent dynamics is transformed into the compact form dynamic linearization data model.Then,control schemes are designed for leaderless and leader-following consensus problems,respectively.Afterwards,based on the squeeze theorem,the stability of the method is proved if the union graph of the communication topologies is strongly connected.Finally,some numerical simulations are given under different packet drop rates,which validates the effectiveness of the proposed method.2.A novel distributed model free adaptive iterative learning formation controller is proposed to solve the formation problem for a class of unknown nonaffine nonlinear heterogeneous multi-agent systems under iteration-varying disturbance.First,the iterative compact form dynamic linearization technique is extended to a disturbance related form.Then,control schemes are designed for measurable and unmeasurable disturbances,respectively.Afterwards,based on the contraction principle,the stability of the method is proved if the communication graph is strongly connected.Finally,numerical simulations show the control performance over different iterations,which verifies the effectiveness of the proposed method.3.A novel distributed model free adaptive containment controller is proposed to address the containment problem for a class of unknown nonaffine nonlinear heterogeneous multi-agent systems.First,compact form dynamic linearization technique is used to transform the agent dynamics into a data model.Then,the local containment tracking error and performance index are defined,and the control schemes are designed for stationary and dynamic leaders,respectively.Afterwards,based on the Lyapunov method,the stability of the method is proved if the communication graph is strongly connected.Finally,simulation results are presented to show the effectiveness of the proposed method.4.A novel model free adaptive iterative learning control scheme is proposed to deal with the perimeter problem of large-scale multi-region traffic networks.First,the single-input and single-output iterative dynamic linearization data model is extended to a multi-input and multi-output form,which is used to describe the state changes of the traffic network in different iterations.Then,the critical value of the fundamental diagram is utilized as the desired target,and the data-driven model free adaptive iterative learning perimeter controller is designed.Considering practical constraints,the control inputs are modified before they are applied to the networks.Finally,five other typical perimeter control methods are selected to compared the proposed method under different testing environment.The results show that the performance of the proposed method is superior than other methods.