Research On Transport Layer Protocol And Resource Management In High-speed Railway Communications

China aims to become a transportation power by first expanding its railway networks.With the development of artificial intelligence(AI),big data and other technologies,high-speed railway transportation is developing in the direction of networking,intelligence,and digitization.In order to ensure the smooth implementation of smart railway,railway communication network needs to meet higher standards.However,the inherent instabilities of high-speed railway network,such as frequent handover,high dynamics and interference,seriously affect the quality of service(Qo S)of high-speed railway network,and the dependability of data transmission of high-speed railway communication system,restricting the safe,reliable and efficient operation of high-speed railway.Therefore,improving the resilience of high-speed railway network and reducing adverse effects caused by the harsh and unstable environment of high-speed railway have become the key problems to be solved in the current high-speed railway network.Resilience is an important characteristic of the network’s normal operation in harsh environment and challenging conditions,and Qo S largely depends on the resilience of the network.The emerging multipath transmission control protocol(MPTCP)and reconfigurable Intelligent surface(RIS)technologies help to enhance the resilience of high-speed railway network.With the help of MPTCP,devices with multiple interfaces such as Wi Fi and 5th generation mobile networks(5G)can transmit data through multiple links at the same time,so as to improve the resource utilization and resilience of high-speed railway network.In addition,RIS can also improve the resilience of high-speed railway network by adding additional reflective transmission links.However,due to the frequent handover of highly dynamic high-speed railway network,the existing congestion control mechanisms of MPTCP are not suitable for high-speed railway network.On the other hand,how to design efficient RIS phase shift in time-varying high-speed railway network to enhance the received signal and reduce interference is also an important but difficult problem.In view of the above challenges,based on ensuring the resilience of high-speed railway network,this thesis carries out the research on MPTCP congestion control mechanism and RIS in high-speed railway environment,so as to improve the transmission rate of high-speed railway and provide technical support for the safe,reliable and efficient operation of smart high-speed railway.The main contributions of this dissertation can be summarized as follows:1)With the aim of solving the problem that MPTCP using the mechanism of "additive increase,multiplicative decrease(AIMD)" can not normally deal with random loss and congestion loss,RVeno congestion control algorithm is proposed,which can effectively couple all subflows by modifying TCP Veno’s AIMD strategy,and evaluate the cause of packet loss(congestion or non-congestion).According to the reason of packet loss,the corresponding congestion window(CWND)adjustment strategy is adopted adaptively.The performance of RVeno is verified based on NS3-DCE platform.The test results show that RVeno can realize load balancing and effectively improve throughput on the premise of maintaining fairness.2)Since the rule-based MPTCP congestion control strategies such as RVeno can not quickly adapt to the highly dynamic high-speed railway network,a congestion control strategy based on deep reinforcement learning(DRL),namely HSR-CC,is proposed.Firstly,to meet the demands of wide coverage and flexibility,MPTCPempowered space-ground integrated network for railway(SGIN-R)architecture is introduced.And then the cross-layer information,i.e.,reference signal received power(RSRP)aided HSR-CC algorithm based on DRL is put forward,which can quickly adapt to the high-speed railway network and alleviate the performance degradation caused by frequent handover.Finally,the effectiveness of HSR-CC algorithm is verified based on NS3-DCE.3)With the aim of alleviating the blockage problem of millimeter wave,the RIS-aided high-speed railway millimeter wave network is studied.With the help of the transmission link reflected by RIS the resilience of high-speed railway network can be improved.To cope with the situation where the hardware of RIS in harsh highspeed railway environment may suffer from the damage unpredictably,a novel DRL framework is proposed to design base station beamforming and RIS phase shift to maximize the spectral efficiency,which combines long short-term memory(LSTM)and deep deterministic policy gradient(DDPG),termed as LSTM-DDPG.The simulation results demonstrate that the proposed LSTM-DDPG scheme outperforms the benchmark schemes in terms of spectral efficiency with comparatively low execution time,which makes real-time decision-making truly viable in the dynamic high-speed railway network.4)The RIS-aided high-speed railway system with interference is investigated.In order to suppress the interference,one RIS is deployed nearby the mobile relay(MR)that is mounted at the top of train.Based on the proposed RIS-aided high-speed railway system model,the RIS phase shift optimization problem is formulated with the objective of maximizing the interference-limited channel capacity subject to unit modulus phase shifting constraints.Due to the time-varying feature of highspeed railway channel,it is challenging to find feasible phase shift design based on the conventional optimization approaches.Therefore,a DRL-based scheme is proposed to efficiently design the RIS phase shift.Specifically,since the phase shifts in the considered high-speed railway system are continuous,DDPG is chosen to develop the algorithm.To facilitate the DDPG-based algorithm,the action space,the state space and reward function are designed.The state space is composed of the time-varying information of the RIS-aided high-speed railway system.The channel capacity is taken as the reward function to train the DDPG-based algorithm.The action space involves the phase shifts,which can be acquired by trial and error.Numerical evaluations are carried out to verify the performance of the RIS-aided high-speed railway communication with external interference.Specifically,the proposed DDPG scheme achieves the higher capacity than the baseline schemes.In addition,the performance gap between the proposed DDPG scheme and the upper bound under the case without external interference is elaborated.The results reveal that by magnificently optimizing the phase shifts,the capacity of RISaided high-speed railway system will improve apparently,and be gradually close to the upper bound with the number of RIS elements increasing.