D29.5 Execution of test for gates installed in Sweden

This deliverable, D29.5 “Execution of Test for Gates Installed in Sweden” addresses the execution of tests for gates installed in Sweden in relation to Task 29.2: Demonstration and evaluation of the installed checkpoints within WP29 – European Railway Checkpoints (ERC). WP29, as part of the Flagship Project FP5-TRANS4M-R, focuses on the development, deployment, and validation of Railway Checkpoints, which are designed to support the automation of freight train transfers at border crossings and other operational stop points.  

This document provides a comprehensive overview of the planning, execution, and analysis of the tests conducted. The purpose of these tests is to assess the performance, reliability, integration, and interoperability of the Railway Checkpoints in Sweden. Two strategic locations have been decided for the installation of gates in Sweden, namely the main gateways to and from Sweden i.e. in vicinity of the Port of Gothenburg, which is Scandinavia’s biggest port, and in Malmö in vicinity of the Öresund’s bridge, connecting Sweden with the European continent.  

The tests in this report are carried out in a comprehensive manner regarding the image analysis for the gate in Gothenburg, and in a slightly less comprehensive manner regarding the newly deployed gate in Malmö. RFID readings of wagon numbers are available for both gates but not yet fully matched with the results of the image analysis. The evaluation of the analysis of the images from the gate in Gothenburg consists of two parts; a manual comparison on a set of images and the data that have been extracted from these images, as well as an automated comparison of data extracted from images and corresponding data from a system called Lupp, which stores data that have been reported by railway undertakings to the rail infrastructure manager in Sweden, Trafikverket. These types of comparisons are made for data that have been produced by two different image analysis systems, one from Indra and the other from Hitachi. Although the results of the comparison are satisfactory on certain detection aspects, further work is needed and currently ongoing in improving the algorithms for image analysis as more images are produced from the installed gates. The gate in Malmö was deployed late March 2026. After that a training period for the gate was launched to train the system to identify all types of EVN, ILU and UN-numbers. At this stage, the system should be considered operational but not yet fully optimized.  

This report also identifies paths for exchange and integration of the data retrieved from the checkpoints with Trafikverket’s data infrastructure and with classical wayside monitoring systems in Sweden.