高铁中基于改进GM(1,1)马尔可夫模型的频谱预测策略

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...

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

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.