Finite-Time Command Filter-Based Adaptive NN Control for MIMO Nonlinearly Parameterized Systems with Time-Varying Input Delay

This paper studies the issue problem of adaptive neural tracking control for multi-input and multi-output (MIMO) nonlinearly parameterized systems with time-varying input delay and unknown disturbances. The integrator with parameters is added to successfully eliminate the influence of time-varying input delay in the process of coordinate changes. The finite-time command filter is adopted to solve the issues of "explosion of complexity " and the error compensation mechanism is introduced to deal with the filtered error. The neural networks (NNs) are applied to approximate the unknown nonlinearly parameterized functions. Then, based on the minimal learning parameters (MLP) algorithm, a new adaptive NN backstepping control scheme is developed and only two adaptive parameters need to be adjusted online in the controller design procedure. Moreover, this proposed method can guarantee that the stability of the overall MIMO closed-loop system and the tracking error converge to be arbitrarily small within a finite time. Finally, the simulation example has evaluated the validity of the control method.