基于递推平方根法的神经网络模型辨识

提出一种基于递推平方根法的神经网络模型辨识方法,对Davidon最小二乘法和阻尼最小二乘法进行了改进,既保持了二者简单易行、收敛性的优点,又能提高精度,减少计算量,适合于应用在非线性系统的辨识和自适应控制中。与常规的Davidon最小二乘法和阻尼最小二乘法进行仿真比较,体现出了这种方法的有效性,尤其是在输入及隐含节点个数较多的情况,其优点比较明显。

Neural Network Model Identification Based Upon Recursive Square Root Algorithm

In this paper ,recursive square root-based learning algorithms for neural network model identification are proposed.As refinde algorithms based on Davidon least squares algorithm and damped least squares,not only do they enjoy the advantages of the two algorithms,simple and convenient to be used,having a fast convergent rate,etc,but they can improve the identification accuracy desired and have less computation.Consequently,they are more suitable for the application in the identification of nonlinear systems and adaptive control.When compared with the Davidon least squares-based algorithm and damped least squares algorighm,impressive simulation results are obtained.