Marking EoA in ETCS and overlaps - some thoughts

Sorry, an engineer trying to second guess what a driver thinks – again! This is triggered by a couple of comments from friends one that in ETCS Level 2 the end of authority needs to be marked at the lineside and the other about the role of overlaps in ETCS.

Before we start, we need to remind ourselves how ETCS works in regard to the extent of a movement authority. The Trackside sends the Onboard information which is related to a balise group known to the Trackside and which the Onboard has already read and reported. The information always includes the distance from the balise group to the end of the authority. It may also include a further distance, beyond the end of authority, to the location of the first hazard or point of collision – we can call that distance an overlap for our purposes.

The Onboard knows how far it thinks it has travelled since it read the balise group and it also has a potential error. That error depends on the design of the odometry system and other factors but is typically 1 or 2% but could be up to 5%. What this means is that if a train considers it has gone 1000m from reading the last balise group it considers that, in reality, it might be anything from 990m to 1010m from the balise group.

Taking a situation (for simplicity) where the overlap is long (say 200m) then the information displayed to the driver telling them where to stop will only be based on the length of the authority. The Onboard will use its estimate of the distance travelled so, based on the DMI information alone the driver would end up stopping the train on approach to, at or even beyond the physical location of the end of authority. This is because the train is just estimating the distance travelled and taking no cognisance of any measurement error.

In reality, driving right to the end of a target braking curve is nigh on impossible and the result is that the drive ends up stopping a short distance before the indicated location. This is due to the characteristics of the brake system, etc, which mean that it is really hard to get right to the target without the brakes being applied by the ETCS. So, the end result is that whilst, theoretically, the driver could be guided, when solely using the DMI, to a location beyond the end of authority, they will generally come to a stand 10-30m before it.

When the Onboard is considering the approach to the end of the overlap, which could be at the end of the authority (so 0m long), then the Onboard adopts a more conservative approach since the end of the overlap represents a hazard which should not be overpassed. For that reason, it considers the potential odometry error and builds it in. If the Trackside declares an authority of 1000m and an overlap of 100m, then the Onboard will assume the odometry is worst case and will supervise an estimated distance of 1000+100m less the possible error – say 11m – and the Onboard will be working to stop the train before it has travelled an estimated 1089m.

The other difference when considering the end of the overlap, known as the Supervised Location, is that the Onboard wants to be sure that even if the driver is incapacitated or driving irrationally that when it decides to intervene it will still stop the train before it has travelled those estimated 1089m. It will, therefore, allow for cutting off the traction power, the potential that the train was still accelerating, the time it takes the brakes to apply and the use of a “guaranteed” or “safe” brake which does not take account of any extra capacity such as regeneration. The brake rate may also be reduced to allow for potential brake defects on the train.

This cautious approach can lead to a passenger train becoming difficult to drive 50 or more metres before the calculated supervised location with the result that if the driver relies solely on the DMI then they would come to a stand about 60m before the physical location.

What is actually displayed to the driver on the DMI depends on the relative positions of the end of authority and the end of the overlap and which is the most restrictive. There are two key pieces of information, the distance to the target in metres (how far the driver could travel) and the current maximum speed the system will permit.

The Onboard is evaluating the required point to brake at the current speed for the end of the authority (based on estimated position) and the required point to brake for the supervised location (based on worst case position and braking). The closest of these points to the current front of the train determines what permitted speed is displayed.

The Onboard also evaluates two distances to the target, one based on the estimated position of the end of the authority and the other the worst case position of the supervised location, both related to the front of the train. It is the shorter of these distances which is displayed.

So, after a lot of technical stuff, let’s turn to what the driver will see in typical situations and consider whether providing a sign or signal at the trackside assists or hinders the driver.

If we take a situation of an end of authority not located at a planned stopping location, such as a platform, and provided for headway/capacity purposes (bit like an old fashioned auto signal), then we can provide an overlap or not. If we don’t provide an overlap in the information from the trackside then the distance to the end of authority will also be treated as the safe distance. This will result in the Onboard using the more conservative brake profile and also making allowance for the potential odometry errors. The net result is that the driver will probably not be able to get closer than about 40-50m of the indicated stopping location on the DMI and that will result in them stopping 50-60m before the actual, designed end of authority.

If one provides a marker at the lineside but the system prevents the driver being able to drive the train to it, then that does not feel right. If one does not provide anything at the lineside then the driver will need to rely on the DMI and just drive inside the braking curves and the fact that the DMI indicates that they could travel a little further probably does not matter because the driver does not need to in order to reach a stopping point or marked location.

If, however, we were to provide an overlap then the braking supervision is more likely to be for the estimated end of authority and the distance will also be estimated. The end result is that the train may come to a stand beyond the physical location of the end of authority without contravening any of the information displayed on the DMI. If we now added a marked location at the trackside w have the risk that a driver legitimately obeying the DMI passes the physical marker – is that a SPAD?

In the situation where the actual stopping location is not critical it appears that the provision of a lineside marker may lead to contradictory information with either the DMI information being more restrictive or inciting the driver to pass the location. In this case it appears that the best way to support the driver is not to provide a marked location.

There will, however, be locations where we want the driver to be able to drive the train to a specific location, say in a platform, and that location will also be at, or close to, the end of authority. In this case a marked location is appropriate and one would expect the driver to take that as their primary information. What you don’t want is then the DMI contradicting what the driver can see with their eyes out of the window nor unduly restricting them. A release speed is often appropriate so that when the train is at a low enough speed on approach to the stopping location the supervision is suspended until the Onboard believes the train is passed the stopping location. The change to the DMI display also provides a visual cue to the driver that they need to get the main information from the trackside.

My conclusion is that, taking a driver centric view, if the location is marked at the trackside then the system must allow the driver to reach the location and that if the stopping location is critical, e.g. in a platform, then it needs to be marked at the lineside. Also, I think, that the DMI information needs to make it clear to the driver whether they are expected to rely on the DMI or to use lineside features for the stopping location and, hence, if one is marking the stopping location it needs to be associated with a release speed to provide that clarity and provide for driveability.

The provision of a release speed generally requires the separation of the end of authority and the supervised location and hence if the stopping location is marked then the movement authority needs to include a distance beyond the end of authority. This can either be by defining the location of the first hazard using a danger point or by defining an overlap.

In addition to, in my view, supporting the driver, this approach also prepares us for moving block where the end of authority could be anywhere and reduces the exposure of our colleagues to track risks since there is less lineside furniture to be installed, maintained and cleaned.

Of course, I am not a driver and it would be good to hear their views. It also requires for all of us, drivers, engineers, operators, etc. to be open minded and to go back to first principles rather than just trying to copy what we have done in the past with a very different technology.