Stewart主动隔振平台的神经网络自适应控制  

Neural network adaptive control of a Stewart mechanism-based active vibration isolation platform

作　　者：马嘉[1] 杨涛[1] 侯增广[1] 谭民[1] 

机构地区：[1]中国科学院自动化研究所复杂系统与智能科学重点实验室,北京100190

出　　处：《控制与决策》2009年第8期1150-1155,共6页Control and Decision

基　　金：国家自然科学基金项目(60635010,60775043)

摘　　要：针对Stewart主动隔振平台,提出一种基于径向基函数(RBF)神经网络的多输入多输出自适应隔振控制方法.考虑外界振动对Stewart主动隔振平台动态特性的影响,建立了隔振平台在工作空间中的动力学模型.推导出RBF神经网络的权值矩阵、高斯基函数中心和宽度的在线自适应调节律,以使神经网络快速逼近系统的非线性动态函数.应用Lyapunov稳定性理论,证明了在扰动力和神经网络逼近误差有界的条件下,闭环控制系统滤波误差和RBF神经网络各调节参数估计误差的一致最终有界.仿真结果表明,该控制方法能有效地抑制不同方向的低频有界振动.For a Stewart mechanism-based active vibration isolation platform, a multiple input-multiple output adaptive vibration isolation controller based on radial basis function network（RBFN） is proposed. Considering vibration effects on the vibration isolation platform, a dynamic model of the platform in the workspace is developed. An online adaptive tuning rule for updating weights, centers and widths of the RBFN is derived to approximate the nonlinear dynamics of the system. By using the Lyapunov synthesis approach, it is proved that the filtered error of the closed-loop system and the parameters of the RBFN are uniformly ultimately bounded in the presence of bounded disturbance force and bounded neural network approximation error. Simulation results show that the developed RBFN controller can effectively attenuates tow frequency vibrations in all six degress of freedom.

关 键 词：自适应控制 径向基神经网络 STEWART平台 主动隔振 

分 类 号：TB535[理学—物理] TP24[理学—声学]