基于Log-Sigmoid型径向基神经网络的雷达辐射源识别

Recogintion of Radar's Radiation Resource with Log-Sigmoid Radial Basis Function Neural Network

提出基于一种新的鲁棒性径向基 (RBF)神经网络的雷达辐射源识别方法。此网络应用Log -Sigmoid函数作为基函数 ,避免了学习过程中的不稳定状态 ,并且比传统的RBF有更好的学习性能和函数逼近能力。并介绍一种新的归一化函数 ,通过归一化函数把不同类型 ,不同量纲的原始评估数值转换到 [- 1,1]区间 ,该效用函数较好地体现了“奖优罚劣”的原则 ,同时又更有利于神经网络的训练。仿真实验证明了该方法的优越性。

A new technique for recogintion of radar's radiation resource with a robust radial basis function(RBF) is presented.The proposed RBF uses Log-Sigmoid function as its basis function that eliminates any risk of instabilities, and it has better learning properties and function approximation capabilities. A new generalizing function is also presented to convert various data with different scale into the interval of .The function carries out the principle to reward the better and punish the worse, and benefits the training ?of neural network.Simulation studies illustrate and support our approach.