基于GAPSO优化的神经网络无线信道参数预测  

Wireless Channel Parameter Prediction by BPNN Based on GAPSO Optimization

作　　者：胡喆馨 卜凡亮[1] 丁丹丹 HU Zhexin;BU Fanliang;DING Dandan(College of Information Technology and Internet Security,People's Public Security University of China,Beijing 100032,China)

机构地区：[1]中国人民公安大学信息网络安全学院,北京100032

出　　处：《无线电工程》2023年第12期2944-2950,共7页Radio Engineering

摘　　要：针对复杂环境中无线信道参数的预测问题展开深入研究,基于反向传播神经网络(Back Propagation Neural Network,BPNN),提出了遗传算法(Genetic Algorithm,GA)与粒子群优化(Particle Swarm Optimization,PSO)算法联合优化的神经网络预测无线信道参数的方法。在利用QuaDriGa平台生成非视距(Non-Line-of-Sight,NLoS)环境下的信道冲激响应(Channel Impulse Response,CIR)的基础上,结合GA在搜索最优解方面的优势与PSO算法加快收敛的特点,利用优化后的BP神经网络模型对相关信道参数进行学习和预测,解决了学习过程中收敛速度慢、预测精确度有限等问题。仿真结果表明,GAPSO-BPNN模型对NLoS环境下的信道参数的预测有较好的效果,能够在未来适应更多的复杂环境。In order to carry out the in-depth study of the prediction problem of wireless channel parameters in complex environment,based on the Back Propagation Neural Network(BPNN),a method of neural network prediction for wireless channel parameters optimized by Genetic Algorithm(GA)and Particle Swarm Optimization(PSO)algorithm is proposed.Firstly,the QuaDriGa platform is used to generate the Channel Impulse Response(CIR)in a Non-Line-of-Sight(NLoS)environment.Then,combining the advantages of GA in searching the optimal solution and the characteristics of accelerated convergence of PSO algorithm,the relevant channel parameters are learned and predicted with the optimized BP neural network model,to solve the problems of slow convergence and limited prediction accuracy in the learning process.The simulation results show that GAPSO-BPNN model has good results on channel parameters prediction in NLoS environment,and can adapt to more complex environment in the future.

关 键 词：信道参数 空间交替广义期望最大化算法 反向传播神经网络 遗传算法 粒子群算法 

分 类 号：TN928[电子电信—通信与信息系统]