THIS year’s ERTMS conference should have been a cause for celebration, as it is now 30 years since the first specifications for ERTMS were agreed. While the rollout of GSM-R, the radio telecommunications element of ERTMS, was rapid, and the development of FRMCS, its successor, is progressing, the deployment of ETCS has been painfully slow and piecemeal.
Kristian Schmidt, land transport director at the Directorate General for Mobility and Transport (DG Move) at the European Commission (EC), told delegates that only 10% of Europe’s TEN-T network representing 12,400km and just 19% of European rolling stock, equating to around 8730 vehicles, has been equipped with ETCS so far. Unsurprisingly, he says the target to install the standardised signalling system on 50,000km of the European Union (EU) core network by 2030 is unlikely to be met.
Small countries have found it much easier than their larger neighbours to install ETCS, although, apart from Belgium and Luxembourg, which have completed national rollouts, deployment has concentrated on parts of the TEN-T network and the newest high-speed lines. This has created ETCS islands and Europe is a long way off from achieving a pan-European network comprising around 100,000km by 2050.
There are some encouraging signs that attitudes to ERTMS are changing, as Enno Wiebe, director general of the European Rail Supply Industry Association (Unife), points out. “ERTMS used to be more confrontational, but now we see less bashing each other and more cooperation,” he says.
France and Germany are a case in point, where there was little appetite to replace established train control systems such as TVM and LZB with ETCS. Only 15% of the French TEN-T network and just 5% of the German network are equipped so far, although the tide appears to be turning.
France is equipping the Paris - Lyon high-speed line with ETCS to reduce headways on this very busy line and plans to deploy ETCS between Marseille and Ventimiglia, followed by Paris - Tours. In Germany, the recent creation of a new ERTMS Coordination Office by the Federal Ministry of Transport (BMV) is encouraging while DB InfraGo awarded a €6.3bn framework contract to four contractors for the supply of digital train control and safety technology, including new digital interlockings, integrated control and operating systems, and the supply of ETCS.
We can’t afford upgrades every five years. We need at least 15 years of stability.
Torsten Lehnert
One of the key objectives for ERTMS is to improve interoperability in Europe. But as Matthias Ruete, the European Commission’s (EC) ERTMS coordinator, points out, only 20 of the 200-plus cross-border rail connections in Europe have ERTMS on both sides of the border.
Ruete suggests creating a European ETCS deployment manager that would act as a central point for EU and national funding for ETCS projects and offer improved coordination across borders.
Even when ETCS has been installed on neighbouring networks, there can still be incompatibility issues. In Luxembourg, 90% of freight operator CFL Cargo’s trains cross a border to Belgium, France or Germany, and even though Luxembourg and Belgium have ETCS, Paul Hodgson, CFL Cargo’s senior business manager reveals that the operator still has to stop at the border for around one hour to reconfigure ETCS.
“This is because the speed profile for freight trains in Belgium is 100km/h, whereas in Luxembourg it is 90km/h,” he says. “The signals in Luxembourg are too close together to allow the speed to be raised to 100km/h.”
Problems
Schmidt lists some of the reasons for the slow rollout of ETCS. These include a lack of standardisation, fragmented national implementation plans, the high cost of trackside upgrades and a slow and complex approval process, and a doubling in the cost of onboard retrofits and upgrading for rolling stock between 2018 and 2023. The cost of retrofitting a locomotive is now as high as €900,000 while upgrades can cost up to €400,000.
Torsten Lehnert, vice-chair of the Association of European Rail Rolling Stock Lessors (AERRL), expands on some of the high costs associated with ETCS. AERRL has 11 members who own around 3000 locomotives, 50% of which are fitted with ETCS. “We are in an overkill of versions and non-compatibilities,” he says.
“You can’t run from one country to another with same version of ETCS, the only exception is the Czech Republic. If you want to adapt a Vectron locomotive fitted with the Austrian version of ETCS 2.3.0d to the same version in Germany the one-off cost is €3.5m. Upgrading a Vectron locomotives from SV2.0 to 2.1 costs on average €250,000 per locomotive. Upgrading a Traxx locomotive from SV 1.1 to SV 2.1 is unaffordable, so the owners are trying to sell them. These upgrades do not provide any additional functionality, they simply mean we can continue to operate the locomotives but with additional costs.
“We can’t afford upgrades every five years,” Lehnert continues. “We need at least 15 years of stability. We can’t have a situation where equipping a short section of line with ETCS, such as opening the new 4.27km Rastatt tunnel in Germany at the end of the year, forces us to have the latest version of ETCS.”
Jean-Philippe Gachot, senior project manager with Swiss Federal Railways (SBB), explains the breakdown of the main cost drivers for ETCS. Design including prototypes accounts for 20-30% of the cost increase, deployment 40-50% and validation and authorisation 20-30%. To reduce costs, he recommends reducing the number of national rules, stabilising specifications for at least 15 years and introducing enhancements in a compatible way. Gachot suggests creating standardised solutions at borders, standardised engineering rules, and modularity.
He also recommends authorising fleets instead of single vehicles, a simplified procedure for massive retrofits, and removing the ETCS test from the authorisation procedure.
“The weight of 30% of ERTMS costs for authorisation should be reduced to a figure similar to the automotive industry which is less than 1%,” he urges.
Gachot also calls for greater prevention of errors. “160 errors have been identified in the Technical Specifications for Interoperability (TSI) in the past 10 years,” he says. “Error corrections are very costly and should not trigger re-authorisation. Every time an error is corrected a new one pops up. We need to stop the errors in the first place.”
Oana Gherghinescu, ERA’s executive director, recommends getting rid of Class B legacy signalling systems as soon as possible, rather than having both Class A (ETCS) and Class B on the same line. “Every kilometre of dual systems adds cost and complexity,” she says.
Nicola Furness, engineering expert CCS, at British infrastructure manager Network Rail, outlines some of the challenges of trying to install ETCS. “You can’t just buy ETCS and plug it in,” she explains. “We have established processes for Class B resignalling projects, but they are not applicable to ETCS. We then end up learning on the job for each project.”
Furness says managing the system once it is installed can also be costly. “Error corrections and cybersecurity updates can lead to higher costs, and minor changes are extremely expensive.”
Go faster
“Rail is doing well, but ERTMS risks becoming the weak link if we don’t accelerate,” Schmidt said, and lists several ways in which the deployment of ETCS could be speeded up.
One idea is to have mature TSIs with harmonisation of ETCS in depth and scope. “We need a freeze on a version after discussion and we need compatibility between countries,” he explains.
Another idea is to reduce the variability of ETCS components by having more standardised products which should help to reduce costs. But as Gherghinescu points out, manufacturers face the problem of having small orders because ETCS projects tend to be relatively limited in scope, which makes it difficult to have long production runs which would reduce costs.
Schmidt says national specifications which increase the cost should be eliminated and suggested introducing a common template for ETCS procurement to help avoid this.
“There is the temptation for clients to add things when they are procuring ETCS which leads to another variant and extra cost,” Gherghinescu explains. “Let’s not be creative. The cost of ETCS is already too high, and we have long delivery times.”
Schmidt wants to accelerate the trackside approval process, which can take up to four years, as Alberto Mazzola, executive director of the Community of European Railway and Infrastructure Companies (CER), points out.
To do this ERA will need more resources. “The European Commission underestimated the resources required for ERA,” Gherghinescu tells IRJ. “We can’t afford to be a system authority with a small team - things are taking longer than necessary, as a result. The EC is being supportive, and it is in their 2026 plan to see what can be strengthened.” ERA currently has an annual budget of €45m, although between a quarter and a third of this comes from the fees it charges for the services it provides. It has a workforce of around 200 people.
“We need to make the case for ERTMS so strong that there is private money for ERTMS, so that we have national, EU and private funding available.”
Oana Gherghinescu, ERA’s executive director
Introducing clear migration plans is vital to accelerate the roll out of ERTMS. Wiebe would like to see a shift from a project-based approach to network-wide deployment.
Several conference speakers said it was difficult to make a business case for investing in ERTMS, which is a reflection of the high cost of components, particularly for equipping locomotives, EMUs and DMUs. This is despite some of the tangible benefits such as greater safety and capacity, and better punctuality.
Belgium decided to deploy ETCS on the national network to dramatically improve safety following a serious fatal accident. Denmark took the plunge because in the early 2000s, a large part of Denmark’s signals and interlockings were 40-60 years old and had become the cause of around 50% of all delays. Within 20 years, 60% of the systems would be obsolete. SNCF has chosen ETCS because it is the only way to increase capacity cost-effectively on its busiest lines. Wiebe says moving block gives a 30% capacity without new infrastructure.
“We need to make the case for ERTMS so strong that there is private money for ERTMS, so that we have national, EU and private funding available,” Gherghinescu says.
The future
Joachim Lücking, head of DG Move’s rail safety and interoperability unit, sets out the EU’s roadmap for future modifications to ERTMS. “The roadmap is based on a strategy adopted in 2024, but it is not set in stone,” he explains.
“We are committed to include the specifications for FRMCS (replacing GSM-R) in the Command and Control Signalling (CCS) TSI,” he says. This will form part of phase 1 of the planned ERTMS modifications in around 2028, together with improved cybersecurity provisions and on-board modularity.
Phase 2 in around 2030 will see the adoption of harmonised engineering rules for ERTMS trackside, a hybrid train detection system to reduce the number of trackside assets which will cut maintenance costs and the full harmonisation of operational rules which will also lower costs.
Phase 3 beyond 2030 will pave the way for the introduction of ATO at grades of automation (GoA) 3 and 4 and an advanced safe train positioning system.
There is clearly greater commitment to ETCS deployment in Europe, and there seems to be more clarity on how to resolve some of the challenges. But infrastructure managers still find it hard to convince their national governments to invest, so Class B systems continue to be installed. Funding for ETCS remains an issue, particularly at a time of economic and security uncertainty. Freight operators find it particularly difficult to justify investing in ETCS as their margins are very tight.
Benoît Gilson, CEO of Infrabel, Belgium’s infrastructure manager, sums up succinctly why Europe needs to commit to ERTMS: “ERTMS is expensive, but it is an investment in the future,” he says. “There are big benefits for safety and capacity, and it secures the railway for the next 100 ways. You need to just start.”
FRMCS proceeds
UNLIKE ERTMS, the development of the Future Radio Mobile Communications System (FRMCS) to replace GSM-R is proceeding on schedule thanks to the skill and the dedication of the UIC team working on the project.
According to Udo Stinner, FRMCS planning manager, FRMCS Functional Interface Specifications (FRS) and the Telecom On-Board Architecture (Toba) FRS have been processed successfully through the Extended ERA Core Team (EECT) and issued by ERA as V2.2.0.
SRS V2.2 was delivered on time at the end of February, fully in line with the initial planning. “This is a major achievement,” Stinner says.
The UIC’s input according to the Version 3 plan will take place in November followed by ERA’s maturity checkpoint in the first half of December. EECT’s V3p period will take place between January and June 2027.
Field tests underway in Germany, the Netherlands, Spain and Sweden will be completed by July 2027.
“We expect to deliver the specifications for FRMCS by the end of 2027 or the beginning of 2028, which is within the timeline specified,” says Begoña Domingo Ortuño, project officer with ERA. This will allow the CCS TSI to be modified to include FRMCS. The first FRMCS products should be available six to 12 months after the new CCS TSI is published.
Dan Mandoc, head of FRMCS, at the UIC, points to the scale of the challenge ahead as up to 120,000km of GSM-R needs to be replaced with FRMCS. Initially GSM-R and FRMCS will operate in parallel.
“Our target is to future proof on-board equipment,” he says. “We are making sure it is backwards compatible.”
Enno Wiebe, director general of Unife, outlines some of the challenges ahead. “Not only is creating the FRMCS specifications a challenge, we must have a roadmap showing how to migrate existing vehicles to FRMCS,” he says. “We must acknowledge the need for FRMCS readiness as a top priority and scale up industrial capacity based on demand.
“At the moment there is no homogeneous, Europe-wide approach to FRMCS migration and GSM-R decommissioning. This lack of alignment will result in long GSM-R/FRMCS coexistence periods, an uncoordinated phase-out, timelines stretching from the early 2030s well into the 2040s and increasing cross-border complexity and coordination risks.
“As a result, defining a FRMCS transition roadmap for existing vehicles is challenging.”
