非高斯噪声中信号的神经网络检测

Neural Detection of Signals in Non-Gaussian Noise

本文研究了非高斯噪声中信号的检测，采用多层感知器神经网络作为检测器。介绍了工作原理，网络结构和训练算法。计算机仿真证明在非高斯噪声条件下神经网络检测器性能优于线性最佳匹配滤波器检测器和局部最佳检测器。

In this paper,the authors study the detection of signals in non-Gaussian noise,and employ a multilayer perceptron neural network as a detector. The authors introduce the operating principle, network structure and training algorithm. By computer simulation,the authors demonstrate that in non-Gaussian noise,neural detectors outperform linear matched filter detectors and locally optimum detectors.