An Innovative System to Detect Obstacles on Rail Tracks

Abstract

Since urbanization continues at a rapid pace, more and more people live in cities. Hence, urban transports systems face an ever increasing ridership demand. As a consequence, many transport operators consider turning conventional (driver-based) systems into automatic operation to increase performance and provide more reliable train service. If converting an existing subway line into fully-automatic operation, the safety of passenger on platforms becomes a vital aspect. An intrusion detection system based on stationary video cameras and intelligent image processing has therefore been developed with dedicated focus on railway safety requirements. Once, the system detects suspicious behavior or a person on the track, central control can be informed and a braking signal can be sent directly to the train via ATP. Since not only the station track is a hazardous zone but passengers standing too close to the platform edge can be at danger as well, a complementary safety function is provided. A 50 cm wide strip at the platform edge is supervised and an audio announcement can be triggered to request people to step back so as to even prevent people from getting too close to the track. This system has been implemented and tested in various installations, e.g., with the Berlin subway operator (BVG) and the subways of Nuremberg and Munich. The system architecture, related safety aspects as well as the operational experiences are presented in this paper.