基于BP神经网络的移动安全性能预测

Mobile Secrecy Performance Prediction based on BP Neural Network

随着第五代移动通信技术的发展,物理层安全成为学术界研究的热点问题.但是由于复杂多变的信道环境,移动物理层安全研究非常复杂,也是一个亟待解决的问题.本文在N-Nakagami信道下,研究了多天线移动协作通信网络的安全性能预测.针对安全中断概率和非零安全容量概率,分别推导了精确闭合表达式.然后基于BP神经网络,提出了一种移动物理层安全性能智能预测方法.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)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 development of the fifth generation mobile communication technology,the physical layer security has become a hot issue of academic researches.Due to the complex and variable channel environment,the mobile physical layer security research is very complicated,and a urgent problem to be solved.In this work,the secrecy performance prediction of the multi-antenna mobile cooperative communication network over N-Nakagami fading channels is investigated.The exact closed-form expressions for the secrecy outage probability and probability of strictly positive secrecy capacity are derived.Then a secrecy performance prediction algorithm based on BP neural network is proposed.Compared to locally weighted linear regression(LWLR),support vector machine(SVM),and extreme learning machine(ELM)methods,the experimental results verify that our secrecy performance prediction algorithm can consistently achieve higher secrecy performance prediction results,which verifies the accuracy of the analytical results.