高铁中基于改进GM(1,1)马尔可夫模型的频谱预测策略  

Spectrum prediction strategy based on improving GM(1,1)Markov model in high-speed railway

作　　者：白天晟 陈永刚[1] BAI Tiancheng;CHEN Yonggang(School of Automatic&Electrical Engineering,Lanzhou Jiaotong University,Lanzhou 730070,P.R.China)

机构地区：[1]兰州交通大学自动化与电气工程学院,兰州730070

出　　处：《重庆邮电大学学报（自然科学版）》2021年第1期34-43,共10页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：国家自然科学基金地区科学基金(61763023)。

摘　　要：针对现有高铁环境中沿线网络复杂且频谱利用率低的问题,将具有人工智能特性的认知基站引入高铁无线通信,并提出一种新的改进灰色GM(1,1)马尔可夫模型对频谱进行预测。与其他方法不同,分别从主用户到来时间及其持续时间两方面进行预测,建立信道的占用/空闲模型。通过新陈代谢GM(1,1)对历史序列的1步预测结果进行对比,得到最佳历史序列个数,并对最佳序列预测值利用二次加权马尔可夫模型进行校正,该校正模型分别对各步长的转移概率和各概率的转移状态进行加权,使其更加适应真实的高铁无线通信场景。通过MATLAB实验仿真,将新的改进GM(1,1)马尔可夫模型与灰色关联度模型的拟合程度和1步预测精度进行对比。结果表明,就时间序列预测而言,该模型对历史序列的拟合程度及1步预测精度更高。因此,该模型能够有效进行频谱预测,提高预测性能。Aiming at the complex network and low spectrum utilization rate in the existing high-speed railway environment,the cognitive base station with artificial intelligence features is introduced into high-speed railway wireless communication,and a new improved grey model(GM(1,1))Markov model is proposed to make spectrum predictions.Different from other methods,the arrival time and duration of the primary user are predicted respectively,and the channel occupancy/idle model is established.By comparing the one-step prediction results of the historical sequence produced in the metabolic GM(1,1),the optimal number of historical sequences is obtained and the optimal sequence prediction value is corrected by the quadratic weighted Markov model.The correction model weights the transition probability of the step size and the transition state of each probability respectively to make it more suitable for real high-speed wireless communication scenarios.The fitting degree and the accuracy of one-step prediction of the new improved GM(1,1)Markov model and the grey correlation model are compared by MATLAB experimental simulation.The results show that,in terms of time series prediction,the model has a higher degree of fitting to the historical sequence and a higher precision in one-step prediction.Therefore,the model can effectively perform spectrum prediction and improve prediction performance.

关 键 词：高速铁路 认知无线电 频谱预测 新陈代谢灰色模型(GM(1 1)) 二次加权马尔可夫 

分 类 号：TN929.5[电子电信—通信与信息系统]