Research On Tracking Control Of Complex System Based On Data Driven

In recent years,tracking control of complex systems is a hot research object in the field of control.Because it is impossible to use accurate and effective models,unmodeled dynamic phenomena widely exist in almost all complex systems.How to design a tracking synchronization algorithm for complex systems with unmodeled dynamic has become an important,interesting and challenging problem.In order to explore this problem,this paper studies the data-driven tracking control of complex systems with observers and the data-driven tracking control of complex systems.The specific research contents of this paper are as follows:(1)A class of complex tracking control systems with observers is studied.Firstly,a data-driven observer is proposed,which uses neural network to estimate the unmodeled dynamics of nodes.Under this observer,the tracking error of the leader node is constructed,and a data-based strategy iterative control scheme is proposed to complete the tracking synchronization of the complex system and realize the state tracking of the leader node.The advantage of the designed observer is that it is based on process data and does not need to know the characteristics of the system dynamic model.The control scheme is composed of action network and critic network.Both action network and critic network are based on neural network.Iterative performance is generated by critic network,and control strategy is generated by action network.In addition,the stability of the observer and the convergence of the controller are analyzed,and the implementation of the control scheme is given.Finally,a simulation example is given to verify the effectiveness of the results.(2)The problem of tracking control for a class of data-driven complex systems is studied.For complex system,by constructing the tracking error of complex system with respect to expected dynamics,the cost function of tracking error of complex system is established,and the cost function is minimized by using adaptive dynamic programming method to obtain optimal control.The strategy iterative convergence analysis method is applied to prove the convergence of the cost function,and the complex neural network is used to realize the convergence iterative algorithm.The complex neural network is composed of action network and critic network.The action network generates control strategy and critic network generates iterative performance index.Finally,the consistency process of the algorithm in complex domain is verified by simulation,and the effectiveness of tracking control is verified.