ONE of the major success stories for rail since the turn of the millennium has been the development and deployment of ERTMS. Intended to facilitate cross-border operation, the combined radio and signalling system is now in use around the world. However, the costs of installing the line-side equipment required by ETCS, the signalling component of ERTMS, can be prohibitive, especially on regional and lower-use lines.
An alternative currently under development that negates the need for line-side equipment is the use of satellite positioning through the Global Navigation Satellite System (GNSS). The GNSS Automated Virtualised Test Environment for Rail (GATE4Rail) programme was launched in 2018 following an open call for proposals from the European Union’s Shift2Rail (S2R) joint venture.
“Europe has a very established solution in ERTMS, but it’s too expensive for regional lines,” says RadioLabs president and GATE4Rail project coordinator, Mr Alessandro Neri. “If you want to lower the costs you have to go to other kinds of technology. GNSS is considered a good alternative, and one of the main points you have to cover to get GNSS into service is the validation and certification of the solution, which the project aimed to achieve.”
A challenge raised by S2R is to limit as much as possible the on-field tests, targeting a zero-test scenario replaced by simulations to reduce the costs, but also to assess the GNSS performance under feared events that are difficult to test since the probability of occurrence is quite low. Specifically, GATE4Rail proposed creating a cloud-based platform to connect existing European ERTMS and ETCS laboratories with GNSS centres, providing a virtualised laboratory test facility that is capable of running simulation scenarios and collecting selected field tests.
The GATE4Rail programme was coordinated by RadioLabs, Italy, with Université Gustave Eiffel (UGE), M3 Systems Belgium SPRL (M3SB), Italian Rail Network (RFI), European Railway Industry Association (Unife), Bureau Veritas Italia (BVI), Centre for Studies and Experimentation of Public Works (Cedex), Engineering and Transport Economics (Ineco), and GNSS Usage innovation and Development of Excellence (Guide) participating in the project.
The Cedex, M3System, Université Gustave Eiffel, Guide, Ineco and RadioLabs laboratories were also involved in the project, with each participant contributing different tools and facilities.
GATE4Rail had two goals: to create a digital model combining both rail and GNSS infrastructure, and to define the testing process framework that tests validity of the model.
“The basic idea was to have a zero on-site certification process, something that you can do in a virtual testbed,” Neri says “You can demonstrate you are complying with the standards in such a way that you can speed up the certification process and reduce the costs.”
The GATE4Rail project was structured into five phases: inception, preliminary design, detailed design and testing, final Proof of Concept (PoC) evaluation, and project support.
The project identified a number of test cases that would benefit from the new platform, including:
- support for new line design engineering
- line engineering validation
- tests for the certification of different components
- support for route compatibility tests, and
- from a GNSS point of view, support for the characterisation of a railway line.
This would allow users to design a new ERTMS line with virtual balise, test a ERTMS line in which the virtual balises are installed, and test the equipment.
Existing methodology and tools have been adopted to simulate GNSS behaviour in different rail-specific scenarios, either under normal operational conditions or injecting GNSS faults at both global and local levels. This also allows very rare faults to be simulated and tested, negating the need for long and expensive on-site measurements.
“We are able to fully characterise the electromagnetic environment and to also generate operational conditions that are not so frequent,” Neri says. “For example, you might have one or two faults with a satellite in a year, so if you want to get a statistical recording your train must travel on the line for 10 years. With GATE4Rail you can assess the overall performance of the ERTMS under hazardous events which are usually not so frequent, but their characterization is a must to guarantee safety. GATE4Rail, we can say, is cable of running a train for a billion kilometres, not just a million kilometres.”
To achieve the second goal, GATE4Rail developed a methodology and tools, assessed by Bureau Veritas as an Independent Notify Body, to automatically update test environments, and provide continuous integration, automated repetition and evaluation through Model-Based System Engineering. This ensures the simulation system can be maintained and scaled, reducing the need for re-assessments and increasing the efficiency of the test environment.
The platform’s subsystem is split into three sets of components: GNSS, comprising the Signal-In-Space (SIS) and augmentation data; train, comprising the Virtual Balise Reader (VBR) and ERTMS/ETCS onboard unit; and trackside, containing the virtual trackside components. Several interfaces between these components were trialled, with the solution with the lowest impact on existing ERTMS/ETCS specifications selected.
In order to accurately simulate the main events that may occur during a real railway operational scenario, GATE4Rail has developed error models for local effects caused by masking and reflecting obstacles that could disturb the GNSS signal reception, such as trees, tunnels, bridges or buildings alongside the track, which are combined with global hazards that have already been defined.
Particularly, common errors observed in environments such as forests, urban areas, or under an open sky were extracted from experimental databases and represented as Gaussian distributions. In addition, errors caused by local obstacles such as tunnels or bridges, which could cause the partial or total loss of GNSS signal, are estimated using record and replay tools and existing databases acquired in previous projects.
The resulting tools allow the GNSS module to randomly generate local errors depending on the type of environment in which the train is travelling. The obtained results proved that the GATE4Rail test platform is a resource to evaluate the concept of Virtual Balise, by providing a useful architecture to validate a GNSS based ERTMS engineering. It allows to testing the equipment under different GNSS conditions, including nominal, real and degraded conditions, in a laboratory environment that can also be used in future to test the introduction of GNSS in ERTMS standards.
The process comprises three stages: preparation, simulation and reporting.
In the first stage, the data is prepared and entered into the system, and the scenario is defined. During the second stage, the ERTMS and GNSS systems and virtual balises are simulated and analysed. In the last phase, a report is generated, which can then be cross-referenced with pre-set Key Performance Indicators (KPI) to determine the success of the simulation.
As well as reducing certification costs and the time to market, the tool can also be used to test the validity of proposals before any money is spent on installing the system. “There’s a trend in Europe to test before you invest,” Neri says.
“Capital expenditure will be lower and you have the possibility to reduce the costs of products and services. It will also allow systems to keep up to date with the latest innovations.”
The future evolution of the GATE4Rail platform will focus on:
- the definition of a service-oriented architecture
- standardisation of the virtual balise reader (VBR) and Augmentation and Integrity Monitoring Network (AIMN) interfaces
- configurable GNSS and ERTMS/ETCS Scenarios by the end-user, and
- full automation of the test execution and results evaluation processes.
One challenge, which has not yet been overcome, is allowing the system to operate in real-time. Neri says that with the existing system, the route of the train is loaded into the system before the simulation is run. However, in future iterations, a closed loop based on faster telecommunications, potentially 5G, could facilitate a real time application.
Neri also foresees the system providing information beyond the GNSS certification process. Further development could include adapting the system to analysing and validating video-based systems, and for use by infrastructure managers to determine the optimal position of balises.
The system could be provided to the industry in two ways: either as a certification as a service, where the consortium can provide the tool, or by selling the software to a certification body such as an infrastructure manager, in which case the IM would provide the facilities such as the computing centre required.
Neri says the positive results generated from the two-year research programme is testament to the successful collaboration of various actors across the industry. “It was really a very good occasion to have such a large number of research centres and other actors working together around Europe,” he says. “It was an excellent opportunity to work together, and the comradery within the group was very strong.”