Neuro observer‐based control of pure feedback MIMO systems with unknown control direction

This study focuses on the problem of neural network (NN)‐based tracking control for a class of uncertain multiple‐input multiple‐output non‐linear systems in pure feedback form. An observer based on K‐filters, is introduced to estimate immeasurable states. In this method, a priori knowledge of the control gain sign is relaxed by using Nussbaum‐type technique. NNs are employed to approximate the unknown non‐linear functions and an adaptive neural output feedback controller is constructed via backstepping technique. The Lyapunov theorem is applied to prove that the overall closed‐loop adaptive control scheme is semi‐globally uniformly ultimately bounded. Finally, simulation results are provided to illustrate the design procedure.