对CMAC神经网络的工作原理进行了简要讨论,针对主动磁轴承具有本质不稳定性和高度非线性动态特性的特点,提出了一种基于CMAC神经网络的非线性控制方案以改善系统的特性.对某单自由度磁轴承的仿真研究表明,CMAC神经网络控制器抗扰动能力强,同时其他控制品质也比传统PID控制器好.
The principle of CMAC neural networks is discussed briefly. Aiming at the highly nonlinear and naturally unstable dynamics of active magnetic bearings (AMB), a nonlinear controller based on CMAC neural networks is proposed to improve system dynamic performance. The simulation results of a single-degree-of-freedom AMB system show that the proposed controller has good disturbance rejection ability; and other control performance is also much superior to the traditional PID controller.