Performance Evaluation Of High-Speed Railway Wireless Network Basing On Stochastic Network Calculus

With the development of the high-speed-railway,higher train speed puts forward more stringent requirements with respect to the type of railway communication service and quality of service under high speed environment,while providing a higher level of convenience for passengers.In order to meet the passengers9 demands of the internet services during the long journey,it is the trend of the time to create the new value-added service when the safety of the railway operation is guaranteed.LTE possesses a higher peak rate,less system delay,flexible bandwidth allocation and higher spectrum efficiency,and hence the railway wireless communication system evolution from GSM-R to LTE-R is an inevitable trend.According to the results coming from the previous measurement of the Internet service,the passengers' data service has the characteristics of the long range dependence and the statistical self-similarity,and it fails to satisfy the memory less characteristic assumed by the Poisson arrival service model.According to the theory of the self-similar network modeling,the passenger service is modeled as FBM model in this paper.This paper focuses on HSR wireless network performance evaluation basing on stochastic network calculus theoretical method,the analysises stochastic network calculus theory and compares the differences between MGF and CCDF stochastic network calculus.This paper summarizes the advantages and the disadvantages of different kinds of stochastic network calculus,and chooses an appropriate method to apply to theoretical analysis.According to the self-similarity of network data flow,the passenger data service is constructed by using the fractal Brownian motion model,and the passenger service data arrival model and random arrival curve model are established by using the effective bandwidth theory.Based on current literatures,two models of high-speed railway mobile relay architecture with relay are proposed.The movement model of high-speed railway and the data process model are established basing on the high-speed railway wireless communication application scenario.Further,the stochastic service curves of two kinds of high-speed railway mobile relay architecture models are derived.Finally,according to the arrival model and the service model,the theoretical delay bound of the high-speed railway wireless network is calculated by stochastic network calculus theory and the theoretical results as well as the simulation results are compared and verified.The innovations of this paper are listed as follows:Firstly,according to the theory of self-similar network modeling,the passenger service is modeled as FBM model and two different methods to generate FBM are compared and analyzed.Then the appropriate method is selected to build the arrival process by the performance of the variance and Hurst parameters.Secondly,the MGF stochastic network calculus and the CCDF stochastic network calculus theory are analyzed and compared.The difficulties in the application of the CCDF stochastic network calculus theory are pointed out.The MGF stochastic network algorithm is used to calculate the backlog and delay bound of passenger data traffic.Thirdly,based on current literatures,two models of high-speed railway mobile relay architecture with relay are proposed.Based on different relay architecture models,the stochastic network calculus theory is used to deduce the delay bounds,and according to the results of the delay bound,this paper gives suggestions about which kind of architecture should be taken.