Advanced Mobile Access Routing for Rail Traffic Management System

Abstract

Railway transportation plays an important role in our daily lives. European Rail Traffic Management System (ERTMS) is the current used railway control protocol, which is the most advanced standard for managing and controlling railway traffic. It has two main components, the European Train Control System (ETCS) and the GSM-R(Global System for Mobile Communications-Railways) for railway operations. There are four main services carried by GSM-R: control information and signaling data exchanging between train and control center, voice communication between train driver and control center staff, regular voice communication between control center staff and railway maintenance workers, and emergency call between train driver and railway maintenance workers. However, exploiting old GSM technology has reached its technological limits (i.e., technological obsolescence). On one hand, there are adjacent channels interference among GSM-R and public mobile networks, and limitation of transmission capacity for GSM-R technology itself. On the other hand, extension to secondary lines and freight trains of safe and efficient traffic management systems, originally designed for high speed trains, requires cost effective solutions for both train localization and communications. For this reason, we propose two alternative solutions to support both signaling data exchanging and voice services. In particular, we use Multipath TCP (MPTCP) solution to sustain the control information and signaling exchanging service, whilst Multipath UDP (MPUDP) solution to support all kinds of voice services. It is worth to mention that both two solutions do not request the dedicate GSM-R networks. The traffics are always delivered over public cellular networks and satellite communication network. In our new MPTCP solution, data packets delivery is supported by both the best effort networks (i.e. Public Land Mobile Networks–PLMNs) and the QoS (Quality of Service) guaranteed network (i.e. satellite network). These two kinds of public networks are cohere by the MPTCP protocol, with the aim of providing handover procedures for seamless connectivity. In practice, once the communication between train and control center is built up, more than one subflows will involve in the data transmission. According to our emulated results, our MPTCP solution can well support the signaling data transmission with high reliability. For the train control voice service, we use multiple UDP links to support the IP based voice packets transmission. UDP packets transported via different paths may have diverse channel delays and packet loss rate. Thus, receiver suffers from a heavy out of order problem. To solve both the out of order and packet losing problems, we introduce RaptorQ protocol in our MPUDP solution. RaptorQ can bring in the redundancy during encode procedure and can execute decode procedure regardless of the packets order. From our test results, we found that RaptorQ rise up packet loss tolerant, and settled the out of order problem, too. Video camera starts to be used in the train positioning service in order to achieve a higher level of positioning accuracy. Meanwhile, with the increasing train speed, real time video monitor for supervising driving conditions, and video call service between driver and control center could be also required in the train control system. Up to now, our architectures only service for the data and voice transmission of railway control system, but in the future, they should bear the railway driving monitor or position reference video transmission and on board public services as well. In order to guarantee the transmitted video quality in control system as well as to achieve a high quality of experience for train passengers, train operators need some methods to monitor the real time transmitting video quality (VQ). In fact, the on board VQ evaluation should request as less reference information from the original video as possible. Thus, we design a no reference VQ metric NRALM for IP based video transmission. It estimates the on board VQ based on only information from received video. Prediction VQ score is computed via pooling all spatial and temporal degradation by means of log-logistic model. To summarize, we propose two communication architectures based on the best effort networks and satellite network, they service for the data and voice transmission, respectively. We give out a no reference video quality assessment method, too. It helps the train operator to achieve a high QoE for the passengers. According to our test results, all of our achievements have outstanding performances