Study On Dynamic Resources Allocation And Queue Management Of Cognitive ENodeB In Railway Environment

Railway transportation as a kind of green and sustainable travel way,it carries the most of the traffic across cities and intercities.In recent years,our country has made great efforts to develop the railway industry.In fact,in the railway communication system based on GSM,the network traffic of system's downlink is small,and the spectrum is idle in most of the time.In addition,the rail lines of our country is “10 vertical and 10 horizontal”,and the utilization of the spectrum is insufficient in many regions,the spectrum is obviously wasted.As a dedicated system of public transportation,although the railway and its dedicated communication network gained rapidly develo pment,but providing a stable and reliable Internet access services for passengers on th e train did not keep up,the issue of bad communication's quality of passengers on train was still outstanding.Therefore,this article focuses on increasing the spectrum utilization of railway dedicated network and improving the quality of communication for passengers on train,the main contributions are as follows:(1)First,for the obviously spectrum waste in the dedicated railway communication system,the cognitive radio technology was used in this thesis to build up a spectrum sharing cognitive radio network(CRN)which based on cognitive LTE-R e Node B(C-e Node B)and vehicle gateway(VG)in the Long Term Evolution for Railway(LTE-R).Next,the functions of the C-e Node B and the VG should have is given,and modified the Round Robin(RR)scheduling scheme.The simulation results show that the proposed spectrum sharing CRN architecture and the modified RR scheduling algorithm can effectively enhance the spectrum utilization of the dedicated railway system and provide an additional communication links for passengers on the train.(2)The queue management strategy based on type of service value priority and the fixed resources allocation strategy based on the available spectrum resources environment map are proposed for C-e Node B,so as to optimize the communication QoS of pass engers on the train.These two strategies are aim ing at reducing the real-tim e services' reception delay onboard.The experimental results show that these two strategies can effectively reduce the reception delay of real-t ime services an d ensure the effectiveness of real-tim e communication.(3)In order to improve the fairness of C-e N ode B' service,a secondary users' queue management strategy based on Kalm an filtering prediction was proposed.Under meeting the real-time services queuing delay,the C-e Node B adjusts the services rates for the realtime services according to the prediction of the length of the real-time services queue.By this way to reduce the queuing delay of the non-real-time service as much as possible.Finally,the simulation reveals that the proposed Kalman filtering prediction method can reduce the queuing delay of the non-real-tim e services an d effectively improve the fairness of the services of the C-e Node B,under satisfying the real-time services QoS.