一种基于DTFEL的非线性自适应逆控制

DTFEL-based nonlinear adaptive inverse control

提出一种基于离散时间反馈误差学习(DTFEL)的两自由度非线性自适应逆控制(AIC)方法,其控制器由动态RBF神经网络(DRBFNN)前馈控制器和参数固定的PD反馈控制器构成.PD控制器用来保证闭环系统稳定,动态RBF神经网络以PD控制器输出和反馈误差的线性组合为学习信号,通过一种改进的NLMS(VS MNLMS)算法在线学习和逼近对象的动态逆,提高反馈控制器的性能.稳定性分析证明了该AIC系统稳定.数字仿真结果表明,该AIC具有良好的自适应能力和鲁棒性,是一种有效的非线性控制方法.

This paper proposes a discrete-time feedback-error-learning （DTFEL） based adaptive inverse control （AIC） strategy with a architecture of two-degree-of-freedom control, the controller consists of a fixed PD feedback controller and a dynamic RBF neural network feedforward controller. PD controller is used to ensure the stability of the closed system. The DRBFNN is driven to approximate the dynamic inverse model of plant by the linear combination of the output of the PD controller and the feedback error signal using the proposed VS MNLMS algorithm. The proposed AIC system is proved to be stable. Practical simulation results show that .the AIC exhibits excellent adaptive ability and robustness performance, and it is an efficient nonlinear control strategy.