论文部分内容阅读
提出了一种新颖的广义径向基函数神经网络模型,其径向基函数(RBF)的形式由生成函数确定.然后,给出了易实现的梯度学习算法,同时为了进一步提高网络的收敛速度和网络性能,又给出了基于卡尔曼滤波的动态学习算法.为了验证网络的学习性能,采用基于卡尔曼滤波算法的新型广义RBF网络预测模型对Mackey-Glass混沌时间序列和Henon映射进行了仿真.结果表明,所提出的新型广义RBF神经网络模型能快速、精确地预测混沌时间序列,是研究复杂非线性动力系统辨识和控制的一种有效方法.
A new model of radial basis function neural network is proposed, whose radial basis function (RBF) form is determined by the generator function, and then an easy-to-implement gradient learning algorithm is given. In order to further improve the convergence speed of the network And network performance, and gives a dynamic learning algorithm based on Kalman filter.In order to verify the learning performance of the network, a new generalized RBF network prediction model based on Kalman filter algorithm is used to simulate Mackey-Glass chaotic time series and Henon’s mapping The results show that the proposed generalized RBF neural network model can predict chaotic time series quickly and accurately, which is an effective method to study the identification and control of complex nonlinear dynamic systems.