摘要
采用基于二层感知网络的判决反馈均衡器结构实现非线性信道均衡，导出了其权值自适应调整的学习算法，给出了一种有效的权值初始化方法。针对一种具有严重码间干扰与有色噪声的非线性信道，分别应用文中提出的二层感知网络判决反馈均衡器和一般的三层感知网络判决反馈均衡器对该信道进行均衡，对其均衡性能进行了计算机模拟，并作了分析和比较。研究表明所提出的二层感知网络判决反馈均衡器无论是收敛速度、误比特率，还是计算复杂度方面都明显优于文献［５］提出的多层感知判决反馈均衡器。
This peper proposes a new approach for nonlinear channel equalization using a two-layer perceptron-based decision feedback equalizer(DFE).An error back-propagation learning algorithm to adjust the weights is derived. A special and effecient weight-initialization method is also given. Both two-layer perceptron DFE and the general MLP DFE are applied to equalize a nonlinear channel with severe intersymbol interference and Gausian colored noise. The equalization performances are obtained and compered by computer simulations. It is shown that the two-layer perceptron-based DFE provides better performances in convergence speed,bit errorrate,and computation complexity than the ordinary three-layer perceptron-based DFE.
出处
《电子科技大学学报》
EI
CAS
CSCD
北大核心
1996年第3期236-240,共5页
Journal of University of Electronic Science and Technology of China
基金
电子部预研基金
关键词
二层感知器
判决反馈均衡器
学习算法
two-layer perceptron
decision feedback equalizer
learing algorithm
weight initalization
