Link State Prediction-Based Reliable Transmission for High-Speed Railway Networks

Due to unpredictable noise and ambient interference along high-speed railways (HSRs), it is challenging to provide reliable Internet services in severe HSR network environments. Most existing research that requires expensive modifications for large-scale already in-used base stations cannot be immediately deployed into the existing HSR systems. In this paper, we propose a quite lightweight but effective solution to improve the Internet experience for HSR passengers. Different from other existing approaches, we employ a data-driven link state prediction (LSP) mechanism for HSR reliable transmission, called LSP4HSR, which directly operates in HSR's on-board routers. In particular, we conduct an extensive measurement of network status on several realistic HSR lines and collect a first-hand dataset in terms of round-trip time and packet loss rate. By analyzing this real dataset, we find that HSR link quality presents obvious two-time-scale variation characteristics. We execute a lot of in-depth studies to explore potential reasons for this interesting phenomenon. Furthermore, based on the two-time-scale Markov chain, we establish an accurate HSR link prediction approach, which brings an LSP-based transmission enhancement mechanism to alleviate the impact from poor link status along HSR lines. Extensive experiments verify that the proposed solution can not only improve the packet transmission reliability in HSR networks but can be also deployed in existing HSR systems quite smoothly and easily.