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. Aiming at Stewart active vibration isolation platform, a multi-input multi-output adaptive vibration isolation control method based on radial basis function (RBF) neural network is proposed. Considering the influence of external vibration on the dynamic characteristics of Stewart active vibration isolation platform, The dynamical model of the platform in the workspace is deduced.The weight matrix of RBF neural network and the on-line adaptive tuning law of Gauss basis function center and width are deduced to make the neural network approach the nonlinear dynamic function of the system quickly.Application of Lyapunov stability Theory, it is proved that under the condition of perturbation force and the approximation error of neural network bounded, the error of the closed-loop control system is consistent with the final boundedness of estimation error of RBF neural network.The simulation results show that this control method can effectively restrain different Direction of low-frequency bounded vibration.