作者: Qichao Zhao , Yan Lin
DOI: 10.1002/RNC.1774
关键词:
摘要: SUMMARY This paper proposes a novel dynamic surface control algorithm for class of uncertain nonlinear systems in completely non-affine pure-feedback form. Instead using the mean value theorem, we construct an affine variable at each design step, and then neural network is employed to deduce virtual signal or actual signal. As result, unknown directions singularity problem raised by theorem circumvented. The proposed scheme able overcome explosion complexity inherent backstepping guarantee tracking performance introducing initialization technique based on error modification. Simulation results are presented demonstrate efficiency scheme. Copyright © 2011 John Wiley & Sons, Ltd.