Research On Multipath Components Clustering For High Speed Railway Wireless Channels Using Machine Learning

With the rapidly development of high-speed railway and the maturity of the high-speed railway technology system,broadband wireless mobile communication has become an indispensable part of high-speed railway services.For the high-speed railway wireless communication system,the propagation characteristics of the wireless channel is an important basis for its design and deployment.Only by fully understanding the wireless channel characteristics of the system,can we adopt various physical layer technologies to fully exploit the channel capacity of the system and further optimize the performance of the system.In the wireless channel,since clustering can describe the distribution characteristics of multipath components of wireless channels more intuitively,in recent years,multipath cluster based wireless channel modeling has been more and more well-known and recognized by researchers.On this basis,this paper combines machine learning theory and radio wave propagation theory to effectively cluster multipath components in high-speed rail wireless channels using correlation clustering algorithms.After the clustering,the multipath components clustering results of power delay profile of continuous time during train operation are tracked.Under these circumstances,the modeling of the time-variant channel of the high-speed railway can be completed accurately,and the data characteristics of the time-variant wireless channel can be deeply studied.Firstly,this paper introduces the commonly used wireless channel clustering algorithm and the optimal classification number evaluation indicator,and compares the advantages and disadvantages of the algorithm according to the characteristics of different algorithms.Secondly,by processing the LTE measured data of the high-speed railway,the high-speed railway wireless channel database is established.Based on the analysis of the existing clustering algorithm and the characteristics of the measured data,these algorithms are combined.And then,the multipath components of the high-speed railway wireless channel are clustered successfully and the Cluster Delay Line model is constructed successfully using above clustering algorithms.During this period,by using different clustering evaluation indicators for the same power delay profile,the superiority and inferior performance comparison between the clustering evaluation indicators was completed,and the appropriate evaluation indicators were selected based on the measured data.According to this evaluation indicator and the number K we got,the clustering of multipath components is completed.In addition,this paper also proposed an automatic clustering method,which is based on the combined application of the aforementioned clustering algorithm.It can implement an automated process from inputting power delay profile data to completing clustering,which can be used for clustering of high-speed railway wireless channel multipath components data and multipath channel modeling in single antenna measurement scenario.And it can also help us to recognize the high-speed railway wireless channel accurately.In addition,based on the above research work,we select a part of LTE measured data of the high-speed railway.After analyzing of each tracking algorithm and the wireless channel measured data in the single-antenna measurement scene,an appropriate tracking algorithm is selected in this paper,and the aforementioned automatic clustering method is improved according to the characteristics of the measured data.Finally,through the combination of the above algorithms,the tracking of the lifetime of multipath clusters during the operation of high-speed trains is completed.By comparing the state of multipath components in the selected time range with the lifetime of multipath clusters in the paper,it can be concluded that the innovation of the algorithm has obtained reasonable and credible results,which is of great significance for accurate research and analysis in time-variant wireless channels of high-speed railway.