基于均匀场理论的CDMA多用户检测的方法

A Mean Field Theory Based Neural Network Approach for Implementing Optimal Multiuser Detection in CDMA System

基于均匀场理论提出了实现码分多址 (CDMA)最佳多用户检测的一种神经网络方法。从理论上证明了该算法的稳定性 ,理论分析和计算机仿真表明 :该算法能有效地抑制噪声干扰、能有效地克服远近效应 ,且能实时求解 。

In this paper, a MFT (Mean Field Theory) based neural network approach for implementing optimal multiuser detection in CDMA system is proposed, and the stability of the algorithm has been proven. Computer simulation by means of Monte Carlo method demonstrates that the algorithm has advantages of better tolerance to coding error and 'far - near' effect over traditional method and decorrelating decision - feedback multiuser detector approach. Other powerful abilities of the approach such as easiness for VLSI implementation and real - time response are also presented. Our approach paves a new way for implementing optimal multiuser detection in CDMA system both for theoretical research and practical applications.