Neural Network-Based Integral Sliding Mode Control for Nonlinear Uncertain Systems
This chapter presents a new integral sliding surface with an adaptive radial basis function (RBF) neural network. In addition to the advantages of no reaching phase and nullifying matched uncertainties, more importantly, it compensates partially for the effects of unmatched uncertainties in the system closed-loop dynamics. Only a part of unmatched uncertainty appears in the resultant system closed-loop dynamics, and thus the system robustness is enhanced. The adaptation law of the RBF network is derived using a defined Lyapunov function. Also based on Lyapunov theory, the switching gain condition is obtained to ensure the system states remaining on the designed sliding surface. Numerical simulations show the effectiveness of the proposed method and improvement against existing methods.
KeywordsRadial Basis Function Slide Mode Control Radial Basis Function Neural Network Sliding Mode Radial Basis Function Network
Unable to display preview. Download preview PDF.
- 1.V. Utkin and J. Shi (1996) Integral sliding mode in systems operating under uncertainty condi-tions. Proceedings of the Conference on Decision and Control, Kobe, Japan, pp. 4591-4596.Google Scholar