基于改进型Hopfield神经网络的盲多用户检测算法  

Blind multi-user detection algorithm based on improved Hopfield neural network

作　　者：王鸿斌[1] 李福平[1] 张立毅[1] 王华奎[1] 

机构地区：[1]太原理工大学信息工程学院

出　　处：《计算机工程与应用》2008年第10期230-233,共4页Computer Engineering and Applications

基　　金：山西省高校高科技开发项目(the Shanxi High- Tech Science and Techology Development Item No.2007149)

摘　　要：基于Hopfield神经网络没有学习规则,不需要训练,也不会自学习,靠Lyapunov函数的设计过程来调节权值的特点,将广义罚函数与Hopfield神经网络的能量函数结合,基于最小平均输出能量准则,构造出更合适的新目标函数,分析讨论了一种实现DS/CDMA盲多用户检测的改进型Hopfield神经网络方法。仿真结果表明,该算法在误码率、抗远近效应方面都有明显的改善。Based on the Hopfield neural network without learning rules,not need training,and not self-learning,to adjust weight by the design process of Lyapunov function,generalized penalty function is combined with the energy function of Hopfield neural network.A more suitable structure of the new objective function is built based on the minimal average output energy norm.An improved Hopfield neural network method of achieving DS/CDMA blind multi-user detection is discussed.Simulation results show that the algorithm is significantly improved in bit error rate and anti-near-far effect.

关 键 词：罚函数 能量函数 目标函数 误码率 远近效应 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]