Optimal State Tracking Control for Linear Discrete-time Systems Via Value Iteration

In this paper, the data-based solution for state tracking control problem of linear discrete-time systems is proposed by value iteration, in which only the system state and control information are required. Instead of directly constructing the error system, the Bellman equation is obtained via constructing an augmentation system. Different from previous researches on state tracking control problem, the Q-learning based on value iteration is used to solve the augmented Bellman equation of state tracking control, which does not require initial allowable control. Finally, the feasibility of proposed algorithm is verified by simulation.