In recent years, a strong push towards driverless mobility solutions can be seen in many transportation sectors including railways. While the European Train Control System already specifies the necessary interfaces to open up the possibility of Automatic Train Operation (ATO) for mainline railway vehicles, required infrastructure-side upgrades of interlocking systems are time- and cost-intensive. Alternatively, a pure vehicle-side Automatic Train Operation solution can be conceptualized that relies on processing the same audio-visual input a human train driver would normally base his decisions on. This would require the vehicle-side detection of track-side railway signals to determine the vehicle’s movement authority and allowed maximum speed. Such a signal detection system could furthermore be employed as an Advanced Driver Assistance System (ADAS) or support autonomous shunting operations. To enable such a system, this paper presents GERALD, a novel dataset for a neural network based detection approach of railway signals. The dataset contains 5000 images from a wide variety of railway scenes as well as annotations for the most common types of German mainline railway signals. The material was gathered using publicly available cab-view recordings uploaded by railway enthusiasts on YouTube. Using a state of the art neural network architecture for evaluation, we notice promising detection accuracies despite GERALD being a comparably small dataset. The dataset is freely available for research and non-commercial purposes at: https://github.com/ifs-rwth-aachen/GERALD

The necessary inspection of railway tracks currently still requires high effort and costs, even though cheap sensors and IoT-capable devices are widely available. Such devices could be installed in regular in-service rail vehicles and potentially monitor and measure track irregularities comparably well. This would enable new possibilities for railway operators as they could provide services that are traditionally executed by network operators. Smartphones are a typical example of these kind of devices. Existing studies already researched upon the usage of smartphones in rail vehicles to potentially measure track irregularities. However, from our knowledge none of the previous studies assessed the quality of smartphone accelerometers. Therefore, we conducted experiments on a shaker test rig with different smartphones to study the frequency response of different devices. This paper presents the results of these experiments and shows that smartphones are in general not suitable to measure track irregularities directly. We show that the quality of the data only allows for monitoring applications with the focus of detecting larger deviations over time. The exact calculation of deviations respectively track irregularities is not viable from our perspective.

For analysing, understanding and predicting the track/train-dynamics in order to develop comfortable and sustainable vehicles a sufficient description of the track course and conditions are key requirements. Not only the track irregularities but also the horizontal curvature affects the vehicle dynamics strongly. Nearly all cities owning light rail systems have gradually established the light rail traffic. By reasons of building density, road transportation infrastructure, and progress of the overall urban planning, the light rail infrastructure was constrained to the pre-existing environment. Thus, the track course is mainly optimized for efficient space use but not for best possible vehicle dynamics. To be able to analyse the track layout of tram networks at a bigger scale, an appropriate methodology that allows acquiring track course data is needed, which is the main objective of this paper. For this purpose, an open data approach was developed by the authors utilizing OpenStreetMap (OSM) to derive the horizontal track curvature based on geodata. This groundbreaking approach improves the state of the art since professional geodetic measurements of light rail tracks are generally rarely publicly available, cost-intensive and their preprocessing can potentially be time consuming. The outcome is a simple, robust, and fast approach that was validated using already existing reference track data which was available to the authors. Additionally, an error estimation of the methodology was carried out. Using a quadratic error function, the median standard deviation of the curvature can be determined and used to rate the exactness of the estimated curvature depending on its magnitude. In this approach, the curvature estimations exactness is generally high for small curve radii and decreases for bigger radii. Therefore it can be concluded that the field of application is especially promising for light rail infrastructure. But also for mainline tracks the new method can be used as a rough estimate, if no curvature data is at hand.