Great Britain’s level crossing safety record is one of the best in the world and overall the risks are well managed. However, level crossings are one of the biggest sources of railway catastrophic risk and every incident has the potential for significant danger to both users of the crossing and trains.

Risk control should, where practicable, be achieved through the removal of level crossings and replacing them with bridges, underpasses or diversions. This is easier said than done, though, as the specification often includes providing access for disabled users and infant buggies. This may require ramped access which, in turn, demands land take and creates a visually intrusive construction. In many cases, there is simply not the land available.

Crossings connect communities, and closing them with diversions will, understandably, meet with local resistance. Where removal is not possible, the risks need to be reduced as far is reasonably practicable, and this is an area where innovative technology at an affordable cost can help to reduce risk.

Risk reduction measures have to take into account all the users of crossings, with ‘equal access for all’ being a requirement. Users of crossings may be mobility impaired (and the population is getting older). They may have hearing and sight impairments and their first language may not be English. Users may be old, young, short or tall, and those riding on mobility scooters and horses will have a different angle of view of crossing signs, signals and approaching trains compared with pedestrians.

AHB crossing with a large skew angle.

AHB crossing with a large skew angle.

Moving to remote operation

The classic highway level crossing originally consisted of gates, controlled by an operator, which were closed to road traffic when a train approached. Such crossings were made safer with the provision of interlockings, so that protecting signals could not be cleared until the road was fully closed and the gates locked.

Similarly, approach locking ensures that the gates cannot be opened until a train has passed over the level crossing, identified by the track circuits having cleared. The crossing operator also has to check that the crossing is clear and that nobody is trapped inside the gates before the signals are cleared to allow a train to use the crossing.

While barriers have now largely replaced the use of gates, they are still found at a number of level crossings.

Remote operation of level crossings was made possible with the introduction of CCTV technology, with an operator still responsible for checking that the crossing was clear. Such crossings are not without risk, though, and the signaller/operator has to be trained and monitored, along with the asset condition, to make sure the safe observation of the crossing is not affected by signaller error (workload), poor equipment (picture quality), or poor contrast as this can make the user’s clothing appear to blend into the background.

There is a limit to the number of crossings that can be operated by one person via CCTV, and the creation of larger signalling centres has required the development of obstacle detection equipment.

The crossing at Ufton Nervet, Berkshire, where seven people died as a result of a collision between a car and a train in 2004, is an example of an AHB crossing where the railway crosses the road at a high skew angle. After four more fatalities, in 2009, 2010, 2012 and 2014, the crossing closed, replaced by a bridge in December 2016.

The crossing at Ufton Nervet, Berkshire, where seven people died as a result of a collision between a car and a train in 2004, is an example of an AHB crossing where the railway crosses the road at a high skew angle. After four more fatalities, in 2009, 2010, 2012 and 2014, the crossing closed, replaced by a bridge in December 2016.

Video analytics

Radar devices which ‘sweep’ the crossing and check whether it is clear have now been installed at many locations, and the concept has been a major success in fully automating the operation of full barrier crossings which are now known as obstacle detected, or simply ‘OD’, crossings.

One obvious limitation to this method of obstacle detection is that the camera or sensor needs to operate reliably in all light conditions, including fog and falling snow. It must also not be ‘fooled’ by harmless obstacles such as paper or cardboard which may be blown onto the crossing. As a result, new methods of obstacle detection are being evaluated for the next generation of OD crossings, including video analytics.

Systems that use video analytics may provide other benefits, such as generating data on the users of the crossing to feed into automatic risk-analysis systems. This will enable the identification of any changes of usage pattern of the crossing being automated, subject to data protection and privacy requirements.

Sunday markets, new housing developments, school bus routes, or even satellite navigation algorithms, can all change the risk profile of a level crossing. Currently, it can be difficult to identify any changes in use without expensive traffic surveys, and even then these will only provide a snapshot in time. The use of ‘big data’ systems is an area which is becoming very important for railways to help manage assets and risks, and to target interventions and resources.

The AOCL crossing at Wraysholme, Cumbria, has been fitted with barriers to prevent accidents such as the fatal one in November, 2008, when a train hit a car that was driven onto the crossing as the train approached.

The AOCL crossing at Wraysholme, Cumbria, has been fitted with barriers to prevent accidents such as the fatal one in November, 2008, when a train hit a car that was driven onto the crossing as the train approached.


The introduction of light emitting diode (LED) technology has improved both the light output and the reliability of lights associated with level crossings. They were first introduced for use as barrier boom lamps, as the vibration when the barriers were raised or lowered affected the life of the traditional incandescent lamps. The benefits of LEDs have also been successfully migrated to the road traffic light signals (Wig-Wags), replacing the traditional 36-Watt incandescent lamps.

The sharp on-off of an LED’s light output, compared with the rise-fall of a traditional incandescent lamp, led to changes to the flasher drive and the proving circuitry of the light, as LEDs require less current to operate. The reliability of LED technology is now so good that there may be more failures caused by proving circuitry rather than by the LED unit itself.

Automatic Half Barrier (AHB) crossings

Since AHB crossings were introduced by British Rail in the 1960s, there have been a number of changes to the design. Currently, they are not used across more than two tracks and the speed of trains over an AHB crossing should not exceed 100 mph.

Trains should not normally arrive at the crossing in less than 27 seconds after the amber lights of the road traffic light signals first show, with at least 95 per cent of trains arriving within 75 seconds and 50 per cent within 50 seconds.

There is no limit to the amount of road traffic but the carriageway on the approaches to the crossing should be sufficiently wide to enable vehicles to pass safely, with the road layout, profile and traffic conditions such that road vehicles are not likely to become grounded or ‘block back’ and obstruct the railway.

A good road profile is particularly important at an AHB as, should the crossing become occupied by a stationary road vehicle, there is no mechanism to detect the crossing is not clear and stop an approaching train. An emergency telephone is provided to alert the signaller if the crossing is occupied, but a train may only be 27 seconds away and with no means of stopping it. This is why only half-barriers are provided, so as to not trap vehicles and pedestrians inside the crossing barriers.

So long as they are used correctly, AHB crossings are efficient and, when compared to other types of crossing, are closed to road traffic for a relatively short period of time. This allows road traffic to keep moving with minimum disruption from the railway.

Unfortunately, AHBs are prone to misuse, with the risk of road traffic ‘weaving’ around the barriers when they are down. Pedestrians approaching the crossing on the right-hand side, when the crossing is in use, will be met with no barrier and only the flashing light and audible alert to stop them. The problem can be worse at ‘skew’ angled crossings with a longer time required to walk over the crossing. AHBs are therefore not suitable where there is a high volume of pedestrian users without other mitigations such as a ‘standing red man’ signal to supplement the road lights.

There are over 400 AHB crossings in use on the Network Rail network and, due to their use on higher-speed lines and potential for misuse, they present the highest risk of all crossing types. So, what can be done to reduce the risk of AHBs, while retaining their benefit of not inconveniencing road users too much?

AHB plus

As long ago as 1983, the Government commissioned a report on pedestrian safety at level crossings due to concerns about their automation. This report made a recommendation concerning the use of pedestrian signals, but the introduction of these has not been widely adopted across the network.

This is now starting to change. Suitable products are now readily available and the introduction of low-current LED signals has assisted their cost effectiveness. The ‘standing red man’ signal is particularly useful in reducing risk when installed for pedestrians approaching an AHB crossing in the opposite direction to road vehicles exiting the crossing, where there is no barrier.

An additional measure for AHBs could be the introduction of programmable red LED road studs across the whole of the carriageway (Fig.1). These would deter vehicles and pedestrians from entering the crossing once a train has passed the strike in point.

A method of reducing the risk of vehicles weaving around the AHB barriers could be to provide a raised ‘median strip’ or central reservation in the area that separates the opposing lanes of road traffic. Such strips have been provided at level crossings in a number of countries around the world. The problem, though, is that a ‘misuser’ could still cross to the other side of the road before reaching the strip, which would also introduce a hazard to cycles and motorbikes, so this option is not favoured by the road authorities.

A more practical way of reducing the risk of weaving at AHBs would be to provide full barriers, along with the use of an obstacle detection device to raise the barrier if someone, or a vehicle, was trapped inside the barriers. This would be known as AHB plus and a number of configurations are currently being evaluated.

The first option would be to ‘stagger’ the position of the additional exit barriers so that a narrow gap was available to pedestrians between the ends of the barriers (Fig.2). A simpler ‘vehicle only’ obstacle detection device would be used to lift the exit barrier if a vehicle was detected when a train approached the crossing.

Another option could be to provide an obstacle detector for both pedestrians and vehicles, which would either prevent the exit barriers closing or lift them if the crossing was occupied (Fig.3). A further safety enhancement could include the provision of a TPWS (train protection and warning system) trigger to stop a train before it reached the occupied crossing. The time factor would need to be calculated for each site, dependent upon the gradient between the TPWS trigger and the crossing. The strike-in point would need to be extended, possibly using an overlay axle counter, to provide (for example) in the order of 54 seconds at 100mph with 12%G braking. An allowance would also need to be included for ‘another train coming’ controls. An indication to advise a driver that his train TPWS had been triggered by a level crossing obstacle detection device would also need to be considered.


Automatic Barrier Crossings Locally Monitored (ABCL) and Automatic Open Crossings Locally Monitored (AOCL) were introduced to automate the operation of crossings on lines with lower line speeds than for an AHB. They are used on lines of no more than a 56mph line speed and ABCL crossings appear, to the road user, similar to an AHB crossing with a single barrier on both sides of the railway. Both AOCL and ABCL crossings are protected by road traffic light signals along with an audible warning for pedestrians.

As with AHB, the barriers on an ABCL only extend across the entrances to the crossing, leaving the exits clear, and are normally initiated automatically by an approaching train. However, unlike an AHB, the operation of the crossing equipment and the absence of an obstruction on the crossing are monitored by the driver of an approaching train, hence the term locally monitored. Train drivers are required to stop their trains short of the crossing unless they have received an indication (in the form of a white light) to confirm that the crossing equipment is functioning correctly and have observed that the crossing is clear.

AOCL crossings are similar but, as the name suggests, they are open, with no barriers and with only the flashing lights and audible alert to stop users entering the crossing. Following a number of serious accidents, with vehicles entering the crossing just as a train was approaching and with no time to stop, AOCLs on the main rail network have been retrofitted with entrance barriers. As far as users are concerned, they now appear to be the same as ABCL crossings, but are known as AOCL+B. The reason for this is because the barrier is a ‘bolt on’ to the existing AOCL circuitry, and there are subtle differences in failure modes and how they operate under local control compared to an ABCL, hence the different designations.

With the lower line speed and local monitoring, along with trains being able to stop if the crossing is occupied by a stationary road vehicle, both ABCL and AOCL+B crossings carry a lower risk than an AHB. However, the risk of road vehicles weaving around the entrance barriers is similar, therefore the addition of exit barriers would further reduce the risk profile.

Such a crossing would be known as an Automatic Full Barrier Crossing Locally Supervised, AFBCL. This is fundamentally a locally monitored crossing with elements of an OD system to determine whether, once the entrance barriers are down, the crossing is clear of standing pedestrians and the lowering of the exit barriers can commence.

The Drivers White Light would only be given once all the barriers were fully down. In the unlikely event of a trapped user (vehicle or pedestrian), the train driver would be able to raise and re-lower the exit barriers using a Drivers Release Unit (DRU).


A new challenge faced by level crossing engineers is integrating the operation of level crossings with ETCS (European Train Control System) Level 2 Baseline 3.

Reducing the variation in road closure times not only minimises the economic impact of unnecessarily delaying road or rail traffic at level crossing, but it also influences risk, as large variations in closure times have been known to increase the risks taken by road users. If the variance in closure times is reduced, impatient drivers are less inclined to ignore, or weave around, barriers.

Systems to provide constant warning times at crossings for large variations of train speeds have had mixed success. ETCS, with its constant ability to report train location and speed, could reduce the amount of variations in the warning times and replace the trackside strike-in equipment. It is an area that requires further development, and one that is very important for Great Britain, given the number of crossings that may exist on routes to be fitted with ETCS.

Level crossings in the future

Autonomous road vehicles are just around the corner (no pun intended) and will be capable of being connected to the environment in which they operate. So, could the automatic level crossing of the future communicate directly with approaching vehicles and warn drivers that a train is approaching? Could the level crossing system of the future actually take control of a road vehicle and bring it to a stop safely before reaching the crossing?

The collection of real time data about crossing use will be key to the management of level crossings, both in asset management and in real time terms. Could ‘big data’ from the road system be linked to the rail traffic management system, such that the railway operation may be modified if, say, children or a slow-moving vehicle approach the crossing?

What is certain is that technology advancements will continue to reduce the risks associated with level crossings.

Read more: Drones – Are they the Future for Track Renewal Surveys?