摘要
在小波神经网络的基础上提出了一种辨识非线性动态系统的方法．该方法有效地将系统辨识所需要的结构形式与多层神经网络及小波基函数所构成的分辨率信息处理过程相结合，建立了从数据到符号的自适应机制．仿真结果表明，该方法具有收敛速度快、逼近精度高、鲁棒性好等优点．
おhis paper presents a method for nonlinear dynamic system identification based on the wavelet neural networks. It integrates the structural form required by system identification with the multiresolution information processing constructed by multilayered neural networks and wavelet base function; the whole adaptive mechanism is also built. The simulation result shows that the method possesses the advantadges such as fast convergence speed, high approximation accuracy and resonable robustness.
出处
《中南工业大学学报》
CSCD
北大核心
1998年第3期280-283,共4页
Journal of Central South University of Technology(Natural Science)
基金
国家教委博士点基金
关键词
系统辨识
小波神经网络
非线性动态系统
system identification
wavelet neural network
nonlinear danamic system
