移动协作通信系统中断概率性能智能预测

OP Performance Intelligent Prediction of Mobile Cooperative Communication System

随着移动用户的数量呈爆炸性增长,第五代移动通信技术得到了迅速发展.但是由于复杂多变的移动通信环境,移动通信系统的性能研究非常复杂.本文在2-Nakagami信道下,基于多发多收(multiple-input multiple-output,MIMO)和混合译码放大转发(hybrid decode-amplify-forward,HDAF)协作通信技术,建立了移动协作通信系统模型,设计了两种发射天线选择(transmit antenna selection,TAS)方案,研究了移动协作通信系统的中断概率(outage probability,OP)性能.针对两种TAS方案,分别推导了系统中断概率的闭合表达式.然后基于BP神经网络,提出了一种移动通信系统中断概率性能智能预测方法.和极限学习机(extreme learning machine,ELM),局部加权线性回归(locally weighted linear regression,LWLR),支持向量机(support vector machine,SVM),广义回归(generalized regression,GR)神经网络,径向基函数(radial basis function,RBF)神经网络等方法进行了比较,仿真结果表明:本文所提出的算法预测性能更好,理论分析的正确性得到了验证.

With the explosive growth of the number of mobile users,the development of the fifth generation mobile communication technology has developed rapidly.Due to the complex and variable channel environment,the research on the performance of mobile communication system is very complicated.In this work,based on multiple-input multiple-output(MIMO)and hybrid decode-amplify-forward(HDAF)cooperative communication technologies,we establish the mobile cooperative communication system model,design two transmit antenna selection(TAS)schemes,and the outage probability(OP)performance over N-Nakagami fading channels is investigated.The exact closed-form expressions for the OP are derived.Then a OP performance prediction algorithm based on back-propagation(BP)neural network is proposed.Compared to locally weighted linear regression(LWLR),support vector machine(SVM),generalized regression(GR)neural network,radial basis function(RBF)neural network,and extreme learning machine(ELM)methods,the experimental results verify that our OP performance prediction algorithm can consistently achieve higher OP performance prediction results,which verifies the accuracy of the analytical results.