Improved I/O Data-Based Model-Free Adaptive Control Technology in MIMO Discrete-Time Systems

Abstract

As one of I/O data-based model-free adaptive controllers, the conventional MFAC controllers only with zero-order output information of the system are widely used around the world. With the NARX model of the discrete-time system, the conventional MFAC controllers are reviewed based on some assumptions together with matrix-inversion-lemma and extreme value theory, parameter estimation method mathematically. On account of one-order output and a new weighted one-step-ahead control input cost function including one-order output, the new D-MFAC controller is obtained and its convergence is improved accordingly. Meanwhile, some simulations are made between the conventional MFAC controller and the D-MFAC controller with the same initial parameters in this paper. In result, the robust is proved and some conclusions are drawn