一种用于GSM系统的神经网络均衡器

A Neural Equalizer for GSM System

在泛欧ＧＳＭ移动通信系统中，由于存在码间干扰，必须采用信道均衡技术。常用的最大似然序列估计（ＭＬＳＥ）均衡器采用维特比算法（ＶＡ），它只能用于具有较小时延扩展的信道。本文给出了一种适用于ＧＳＭ系统的部分连接神经网络均衡器，其复杂度与信道的时延扩展成正比，特别适用于具有较大时延扩展的恶劣信道。为了消除神经网络局部极小点引起的性能恶化，我们采用了均场退火技术。本文提出的均衡器既适于用模拟集成电路实现，又适于用数字集成电路实现。计算机仿真表明本文的方法具有比最常用的１６状态和３２状态ＶＡ更好的性能。

Channel equalization is essential in the Pan-European GSM mobile communication system. The Maximum Likelihood Sequence Estimation (MLSE) using the Viterbi Algorithm (VA) is commonly recommended for the equalization, which can only accommodate the channels with limited time delay spread. This paper presents a partially connected neural equalizer for the GSM system, in which the complexity is linearly proportional to the time delay spread. For the purpose of effectively eliminating the performance degradation due to existence Of local minima, the Mean-Field-Annealing (MFA) technique is introduced. The present approach is well suited for digital or analog VLSI implementations and can converge in a very short time period. Computation simulations show that our approach can provide performance improvement over 16-state and 32-state VA's which are usually suggested for practical applications.