Output feedback adaptive neural control for a class of non-affine non-linear systems with a dynamic gain observer

Output feedback adaptive neural control is investigated for non-affine non-linear systems with zero dynamics using implicit function theorem, mean value theorem and neural network (NN) parametrisation by exploiting the explicit Lipschitz property of radial basis function NNs for function approximation. The control approach developed is based on non-separation principle design. A new dynamic gain observer is introduced to estimate the unmeasurable states of the system. The observer gain and the neural controller are simultaneously tuned according to output tracking error. With the universal approximation property of NN and the simultaneous parametrisation both for the NN approximation and the controller, restrictive conditions, such as Lipschitz assumption, strictly positive realness condition and contracting assumption are not required. Semi-globally uniformly ultimate boundedness for the steady-state and transient performance is guaranteed, and simulation results demonstrated the effectiveness of the proposed scheme.