论文部分内容阅读
针对伺服系统的系统参数摄动和非线性动态摩擦补偿问题,提出基于递归神经网络(RNN)的自适应反步控制(RNABC)系统设计方法。RNABC系统由反步控制器和鲁棒控制器组成,反步控制器包含RNN不确定观测器,鲁棒控制器则用来消除由于引入不确定观测器而带来的逼近误差。由于自适应反步控制的自适应律源于Lyapunoy函数的,因此系统的稳定性得到了保证。仿真结果表明,对于系统参数摄动和非线性摩擦干扰RNABC能使伺服系统具有很好的跟踪性能。
Aiming at the problem of system parameter perturbation and nonlinear dynamic friction compensation of servo system, an adaptive backstepping control (RNABC) system design method based on recurrent neural network (RNN) is proposed. The RNABC system consists of a backstepping controller and a robust controller. The backstepping controller includes an RNN uncertain observer. The robust controller is used to eliminate the approximation error due to the introduction of an uncertain observer. Since the adaptive law of adaptive backstepping control is derived from the Lyapunoy function, the stability of the system is guaranteed. The simulation results show that RNABC can make the servo system have good tracking performance for system parameters perturbation and nonlinear friction interference.