Reinforcement Learning-Based Adaptive Output-Feedback Control for Discrete-Time Strict-Feedback Nonlinear Systems

2022 
In recent years, reinforcement learning has been widely concerned in the field of adaptive optimal control. In this paper, a reinforcement learning-based output-feedback control scheme is presented for a class of discrete-time strict-feedback systems. Firstly, a novel adaptive neural network-based state observer is constructed and the boundedness of the state estimation errors as well as the weight errors of observer neural networks are guaranteed. Then, a design of critic-action-based controller is implemented. To be specific, critic and action neural networks are adopted to obtain the optimal update law for the controller in a online manner. Variable substitution technique and estimated states are employed in the backstepping procedure instead of the n-step ahead predictor. Therefore, the n-step time delays are avoided during controller implementation. The proposed scheme ensure all the signals of the closed-loop system are uniformly ultimately bounded. Finally, simulation studies are provided to demonstrate the presented scheme.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    17
    References
    0
    Citations
    NaN
    KQI
    []