Requirements for Safety Relevant Positioning Applications in Rail Traffic

During the last years several research projects on the development of positioning systems for trains by means of satel-lite based systems were carried out. Some of these systems, e.g. SATNAB, LOCOPROL and APOLO, were presented in form of demonstrators and validated against their suitability. The Institute of Transportation Systems (IFS) of the German Aerospace Center (DLR) in Braunschweig gives special attention to basic research in this field, especially on the vehicle autonomous positioning with view to operational and safety relevant issues. The positioning of vehicles in rail traffic plays a significant role in the system safety, since the position information is used for the train separation and train control. The basis for giving a route to a train is the safe information that the track is unoccupied. This again needs an integrity check and safe train end identification on board of the train. There-fore, the generation of position information is a safety relevant application. Safety critical requirements arising from the criticality of the information are accuracy, reliability, integrity and availability of the positioning information that are to be demonstrated according to the new CENLEC standards EN 50126 [EN126] and 50129 [EN129] with the safety integrity level (SIL) 4, which is the highest level used in the area of railways. From the operation mode of the track result further requirements, e.g. precision of positioning information of 1 to 2 metres to receive a track selective train position, or the real-time position calculation for the detection of the safe train end in the “Moving Block” mode. However, the characteristics of the satellite based positioning systems (GNSS ), such as GPS or the future European system Galileo, bear significant problems regarding the availability of the satellite signals, and, resulting from that, problems in the quality of the calculated information. This is especially problematic in tunnels, roofed railway stations, canyons and woodlands, as the signals are shadowed and therefore can’t be received, at least not in sufficient signal strength. The same applies to trains in urban areas, e.g. in stations of big cities, were the signals are shadowed by buildings and the signal propagation delay is extended by refection or multi path propagation. In these cases the position information, gathered from these signals, is not equal to the real train position but rather a random position. This deviation can reach a safety critical amount. The effects of multi path propagation or complete shadowing in urban areas and their relevance to the development of a safety relevant satellite bases positioning system haven’t been taken into account in the so far investigated systems. A significant need for research has been identified in this area. From the stated requirements of the rail system and the characteristics of GNSS emanates the conclusion, that the position of a train can’t be calculated sufficiently on the basis of satellite based positioning systems alone. Hence, further data has to be included in the calculation of positioning information. Several approaches for the solution of this issue exist already. Some of these approaches shall be discussed below