An indirect data-driven method for trajectory tracking control of a class of nonlinear discrete-time systems

This paper presents an indirect data-driven method for the trajectory tracking control problem of a class of nonlinear discrete-time systems, which have unknown dynamics. This method first establishes an approximate model of the controlled object using historical I/O data and neural network; then, designs and adjusts the feedback gain matrix online using measured output data and previous estimates. This is an adaptive control process of prediction, estimation, and adjustment, which needs to solve some nonlinear optimization problems online, can overcome the adverse effects of the modeling errors caused by neural networks, and is the key to making the system output asymptotically track the reference trajectory. The convergence analysis and simulation results demonstrate the effectiveness and feasibility of the presented method. In addition, based on Lagrange's mean value theorem, we also give an online linearization technique which is applicable to nonlinear discrete-time systems, whose dynamic models have continuous partial derivatives with respect to the input and the output.