基于神经预测的模糊主动隔振控制技术研究

摘要

将基于神经网络预测的模糊控制技术应用于静电悬浮加速度计地面实验装置主动隔振系统.神经网络预测模型可以根据当前时刻控制器与隔振平台的输出,能够对下一时刻隔振平台的运动状态提前作出预测,并将这一状态反馈输入到模糊控制器,能够使模糊控制器提前作出判断对隔振平台进行控制.研究结果表明该算法不仅可以限制控制信号的振荡,而且有利于抑制超调,对0.1Hz以上的低频激扰具有良好的隔振效果.%The paper applies neural networks forecast based fuzzy control technology to electrostatic levitation accelerometer ground experiment device active vibration isolation system. The neural networks forecast model refers to present controller and vibration platform output to forecast the next movement status of the vibration isolation platform; then send the status feedback to the fuzzy controller to enable it to make judgements in advance to control the vibration isolation platform. Study shows the algorithm not only can limit the control signal vibration, but also is good at restraining excessive frequency and isolating vibration for low-frequency vibration signals which are above 0.1 Hz.