HomeDigital RailwayERTMS Level 3 - A possible way forward

ERTMS Level 3 – A possible way forward

Listen to this article

The global debate on the merits and problems associated with ERTMS continues unabated. Good progress has been made in recent times with the signing in Autumn 2016 of the Memorandum of Understanding concerning the co-operation for the deployment of the European Rail Traffic Management System. This will ensure the regulatory and design authorities work together and that the latest version of the specification is confirmed as the standard with no deviation from this for new ERTMS projects.

It has taken around 20 years to reach this stage, which says much for the diversity of signalling principles and operating rules across the EU member states and other adjacent countries, and the ensuing difficulty in reaching agreement.

There are many parties involved in this: the European Commission, the European Union Agency for Railways (formerly the European Railway Agency – ERA) and the European Rail Sector Association are the three main ones but the latter comprises the CER, EIM, ERFA, ERTMS Users Group, GSM-R Industry Group, UIC, UNIFE and UNISIG. One can only hazard a guess at the number of meetings and the time expended to arrive at a consensus. Nonetheless, an objective has been achieved in that the standards for ERTMS Level 2 are effectively set in stone.

However, it is not the end of the road as, whilst ERTMS Level 2 brings considerable benefit for interoperability and some capacity gains, it was always envisaged that Level 3 would be the ultimate goal as this offers significant cost savings for infrastructure equipment. Predictions in the mid-1990s that the technology would soon be available proved to be a pipe dream.

So what is ERTMS Level 3 all about and why has it proved so difficult to achieve?

The Level 3 concept

In short, ERTMS Level 3 has two main features over and above Level 2. It facilitates much closer headways by the opportunity of adopting moving block, meaning that trains in close succession can close up particularly at lower speeds, and it allows for the removal of track-based train detection equipment in the form of track circuits or axle counters, which should reduce capital and maintenance costs and improve reliability.

When first conceived, it was the latter of these two that dominated the thinking. In those days, rail was seen as a declining industry and anything that could reduce the cost of operation was to be welcomed. Nowadays, with rail patronage continually increasing year by year, it is getting additional capacity that is the driving force.

So why has Level 3 made such little progress in the intervening period? A recent IRSE lecture, given by Nicola Furness from Network Rail and Henri van Houten and Martin Bartholomew from ProRail in the Netherlands, attempted to answer these questions and included a pragmatic solution to overcoming the fundamental problems that exist.

Level 3 ETCS (the signalling element within ERTMS) is based on a total radio solution. A train’s position is reported back to the RBC (Radio Block Centre) at least every five seconds. This information is based upon the data obtained from a series of track-mounted eurobalises (radio beacons) provided at intervals dependent on the positional accuracy needed (for instance, where a precision stop is required). The position reference obtained is then incremented by accurate train-borne odometry that calculates the distance travelled from the last balise.

This constant updating of position and speed allows following trains to run closer to the one in front by adjustment of the MA (Movement Authority) information displayed to the driver. The potential gain in capacity is significant.

So what is the problem?

Why has Level 3 not yet been developed into a standard way of working. There are a number of issues but the two main factors are that the train has to be proven as complete – that it has not become uncoupled en route with part of the train left behind – and the radio system has to be completely reliable.

For modern passenger trains (usually multiple units), a train data bus exists down its complete length to facilitate brake and traction demands, communication systems and general train condition monitoring. Thus, if a separation were to occur, it would be immediately obvious. Even older passenger trains are likely to have electrical connections down the train that fulfil the same objective.

On freight trains however, no such train integrity exists and, although a broken brake pipe connection will stop the rear wagons, it is possible that the locomotive’s compressor could overcome the resultant air leak so the front part of the train would not be affected and the driver would be unaware that his train had divided. In such a situation, track circuits or axle counters will detect that the train is incomplete and prevent a movement authority of any type (including lineside signals) from being given to a following train. Without an independent train detection system, a different form of Train Integrity Monitoring (TIM) has to be part of a Level 3 application.

If communication is lost because of a train radio failure or the radio network has become defective, possibly through external interference, then the position and speed messages every five seconds cannot be given and the trains will stop, with no easy means of recovery.

Various solutions to these two problems have been put forward but none have proved to be operationally acceptable or have the necessary safety integrity and thus obtaining safety approval would inevitably be difficult.

Hybrid Level 3

The solution now being put forward aims to get around these two problems and allow the increase in capacity that is so urgently needed. Known as Hybrid Level 3, it has been in development since 2013 as a joint effort by Network Rail and ProRail with Alstom and Bombardier both supplying equipment that demonstrated the feasibility. The stage has been reached whereby a potential application to the ‘real’ railway can be considered.

The crux of the proposed system is to retain any existing track-circuit or axle-counter sections and to then create ‘virtual blocks’ as sub sections within these. The operation would be:

  • A train equipped for ETCS Level 3 operation would receive an MA allowing it forward into the block section which, if no other train is preceding it, might be to the end of the section or even beyond.
  • A following train that is also equipped for Level 3 operation would receive an MA to enter the same section with an MA to the limit of a safe stopping distance of the first train, taking into account the distance and speed of both trains. If speeds were low, then the second train could close up on the first under moving-block principles.
  • Any subsequent train also equipped for Level 3 operation would follow in the same way.
  • If a train only equipped for Level 2 operation were to approach the section, it would not receive an MA until all preceding trains had cleared the track circuit or axle counter section. Once this has occurred, the train would receive an MA to the end of the track- circuit or axle-counter section. Any following train would not receive an MA until the Level 2 train had cleared the complete section.

It follows that any train not equipped for either ETCS Level 2 or 3 operation would not be permitted to run on this particular route unless, in addition to track circuits or axle counters, lineside signals are retained. This is similar to the situation today for lines being considered for Level 2 introduction whereby, unless all items of rolling stock using the line are equipped, conventional signals have to be kept in a so- called ‘overlay’ mode.

Technical implications

The train equipment for Level 2 or Level 3 operation is virtually identical, other than a Level 3 train has to incorporate a periodic TIM data signal that the train is complete. The display of the Movement Authority and ancillary information is the same.

The infrastructure will require some development and addition. Firstly, the balise positioning and track-circuit/axle-counter section lengths should be aligned with each other to ensure that positional information, as displayed to the signaller, will be the same regardless of the source. This may not be the ideal situation, since the track circuits and axle counters will be unchanged from whatever existed hitherto. However, it would be a small price to pay compared with the advantages to be gained. The RBC will need to have a ‘bolt on’ Virtual Block Detector, not only to permit the authorisation of the relevant MAs, but also to distinguish to the system which trains are operating in either Level 2 or 3.

The Network Rail/Prorail team has devised a set of conditions that would show the signaller the status of each section and sub-section. The latter would normally either display Occupied or Unoccupied. However, there may be conditions where the status is uncertain and two further conditions are foreseen – ‘Ambiguous’, which means a train is present but its status is not known and ‘Unknown’, where the occupation of a sub section is not proven. A total of over 100 scenarios are being thought up and tested which include all kinds of failure conditions.

GSM-R (and NRN) installation on the Tornado steam locomotive.

The fact that existing ERTMS rules remain largely unchanged, and the type and use of existing train detection equipment remains the same, should mean that preparing the necessary safety cases and having them approved will be relatively straightforward.

Progress to date

Speed is of the essence since capacity gains are required urgently, as Nicola said at the start. However, even with this in mind, it will take time to get the full operational scenarios sorted out together with associated testing and progression of the dreaded approval process. The programme as currently seen is:

  • September 2015 – operational principles established;
  • March 2016 – validation of Hybrid principles in a Siemens laboratory;
  • March 2017 – establish Hybrid principles as a European standard;
  • Late 2017 – operational field trials to commence involving all ETCS suppliers, probably using the Hertford Loop test site;
  • Early 2018 – an early deployment trial on a chosen route;
  • 2019 to 2030 – virtual blocks established as a common feature across Europe, possibly combining this with deployment of a radio system to replace GSM-R.

Ongoing challenges

There can be little doubt that the Hybrid Level 3 concept has a potentially significant impact. Whilst pragmatic, it surely must not detract from efforts to develop a true form of ETCS Level 3. Even though the ongoing use of existing track circuits and axle counters means no additional capital expenditure on these items, they still have to be maintained and eventually updated, which will cost money.

The current ERTMS Command and Control TSI already embraces Level 3, and this must not be forgotten. It would be all too easy to forget the end goal and the reduction in trackside infrastructure with associated cost savings. Many existing rail routes do not carry freight trains and have modern passenger stock – could these not be equipped for true Level 3 operation straight away?

The ‘buy in’ from suppliers needs to be assured. Many of these depend on the sale of products such as track circuits and axle counters as a main income stream and the business model will be adversely affected if this is lost. Some alternative business model may need to be thought through.

Neither does the Hybrid solution do anything to negate the impact of radio failures. This requires urgent consideration as it is a problem that equally affects Level 2 unless lineside signals are retained. Another article as to solutions to this problem will be forthcoming shortly, watch this space.

Finally as one contributor asked, is it Hybrid Level 3 or actually Enhanced Level 2? At the end of the day, providing the capacity gains prove to be real, does it really matter?

Written by Clive Kessell


  1. In the supermarket, all the Alcohol and most of everything else has an RFID tag on it, which can be interrogated locally by a sensor revealing the item’s identity and location. Curious that each and every bogie, or vehicle, cannot be fitted with such RFID tag to identify it’s proximity to a balise and to eliminate the need for a Train integrity monitoring system.


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.