Adaptive control of non-affine nonlinear systems using radial basis function neural network

This paper proposes an adaptive controller with Gaussian radial base function neural network (RBFN). The controller is for a class of non-affine nonlinear systems with ill-defined mathematical model. It can work in conjunction with another continuous controller such as a PID controller to improve the performance. Based on Lyapunov's stability theorem, the adaptation laws are conceived for the parameters of the RBFN, including the output weights, the centers, and the variances of the Gaussian radial functions. A bounding control is also developed to help for stability. The effectiveness of the controller is illustrated on a simulation example of a continuously stirred tank reactor (CSTR).