基于RPROP神经网络的多模盲均衡算法研究  

Research on Multi-modulus Blind Equalization Algorithm Based on RPROP Neural Network

作　　者：吕大千 何俊[1] 李柔刚[1] 

机构地区：[1]电子工程学院,合肥230037

出　　处：《舰船电子工程》2015年第12期44-46,158,共4页Ship Electronic Engineering

摘　　要：针对传统盲均衡算法运算量大、收敛速度慢的问题,结合RPROP神经网络算法运算量小、速度快的优点,提出了一种基于RPROP神经网络的多模盲均衡算法。运用Matlab软件,以16QAM、16PSK信号为输入源,对该算法均衡效果进行仿真研究,仿真结果表明该算法在有效减少码间串扰的同时,缩短了收敛时间,达到了预期效果。Aiming at the problem of traditional blind equalization algrithm＇s high computation and slow convergence, considering the advantages of resilient back-propagation（RPROP） Neural Network＇s low computation and quick convergence, a kind of Multi-modulus Blind Equalization algorithm based on RPROP neural network is proposed. By using Matlab, consid- ering 16QAM and 16PSK signals as input source, the equalization＇s effects are studied. The result of simulation indicates that this algorithm can efficiently eliminate the intersymbol interference and reduce time of convergence. And it meets the ex- pectations.

关 键 词：盲均衡 多模算法 RPROP 神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]