超磁致伸缩致动器的小脑神经网络前馈逆补偿-模糊PID控制

针对超磁致伸缩致动器(GMA)在精密致动控制中存在的迟滞和位移非线性,提出了小脑神经网络(CMAC)前馈逆补偿结合模糊PID控制的新策略。通过小脑神经网络(CMAC)学习获得超磁致伸缩致动器动态逆模型用于对超磁致伸缩致动器迟滞非线性进行补偿;利用模糊PID控制降低小脑神经网络(CMAC)学习时的误差和抑制扰动,提高系统的跟踪控制性能,从而实现超磁致伸缩致动器的精密致动控制。仿真和实验结果表明:所采用的控制策略有效地消除了迟滞非线性的影响,系统的跟踪误差降低到了5%以下,而位移跟踪误差均方差仅为0.58。此外,这种策略的特点是学习和控制同时进行,控制系统能够适应被控对象动态特性的变化,使系统具有较强的鲁棒性,同时也能够有效地抑制外界的干扰,提升系统的自适应控制性能。

CMAC feedforward inverse compensation-fuzzy PID control for giant magnetostrictive actuator

Giant magnetostrictive actuators show hysteresis and displacement nonlinear characteristics when they are used in the field of precision actuation control.To overcome the shortcomings, a kind of control strategy combined with the Cerebellar Model Articulation Controller (CMAC)feedforward inverse compensation and the fuzzy PID control was presented. A dynamic inverse model of the GMA was established with the CMAC on-line learning method to compensate the hysteresis nonlinearity of the GMA. A fuzzy PID controller was introduced in GMA to achieve the precision control of the GMA. Using this controller, the learning error of CMAC was decreased, the disturbance was eliminated and the tracking control performance of system was improved. The results came from simulation and experiments show that the control method can effectively reduce the hysteresis error and the tracking error of the system is less than 5% and the displacement tracking error is 0.58 (mean square error). In addition, this strategy is characterized by learning and controling at the same time, so that the control system adapts to the changes in the dynamic characteristics of controlled object and has a stronger robustness.Meanwhile, it eliminates external interference effectively and improves its adaptive control performance.