Level crossings which are monitored and operated by obstacle detection (OD) technology are now well established and in use on rail networks throughout the world. Technology components and detection techniques are improving all the time, so what do the systems need to do and what are the detection methods available?
An ideal obstacle detection system needs to provide a safety integrity no worse, and ideally better, than a manually operated crossing, cause no or minimal delays to trains due to equipment failure or false detections, be affordable in terms of whole life costs, operate in all weather and temperatures, and be practical to use and maintain.
The detection system is required to confirm that a crossing is not occupied by a person (including small children or someone who may have fallen over) or by any object that may cause damage to a moving train. Separate technology systems are required to confirm that the crossing is closed by barriers or gates and, only once the detection system has again confirmed the crossing is clear, is a train allowed to proceed across the crossing. This could be achieved by clearing the protecting signals or, on some rail networks, by using a direct communication connection to the train.
The OD system needs to confirm that a person is not trapped just inside the protecting barrier and should not be confused by non-harmful items such as umbrellas (not held by a person), cardboard boxes, newspapers, fog, falling snow or heavy rain. The detection system will be exposed to electrical interference from traction and power systems, as well as dust and dirt from passing trains. It must not interfere with any train signalling or communications system and needs to comply with all relevant electromagnetic compatibility regulations.
These requirements are often in conflict with one another. For example, it might be possible for a particular system to offer good safety benefits, but only at the expense of causing significant operational delay. The obstacle detection system must also interface with the existing railway infrastructure, and not effect either the rolling stock or the operational procedures.
An OD system may use one or several of the different types of detection, for example the first generation of Network Rail OD crossings uses both radar and laser image detection and ranging (LiDAR).
When OD crossings were first introduced on the British rail network, a report commissioned by RSSB identified a number of technology options, which included some of the following.
Video and thermal imaging
A manually operated crossing can be confirmed clear by either direct observation or by a competent operator using CCTV. So, could imaging technology automate the human checking of a crossing being clear?
An obvious issue with using video technology is what happens at night or during fog when it is difficult to see, with an automated system unlikely to be as sensitive to difficult light conditions as a manual operator. The crossing may therefore need the same or a higher illumination level than a manual crossing, although even this would be of little use in foggy conditions.
The cameras would be monitored via software algorithms with pixel-by-pixel analysis in order to identify objects or persons. Video imaging systems, however, cannot easily distinguish the mass or material properties of an object, so cannot determine the importance or relative threat of an object. Therefore, for example, a cardboard box or newspaper could be confused with a small child. Other difficulties are being able identify stationary objects, and the boundaries of the crossing so that any movements just outside of the crossing are not confused with an object or person on the crossing.
Thermal imaging cameras can overcome some of these limitations because they create a crisp image based on subtle temperature differences and are not affected by environmental challenges, such as total darkness, smoke or fog. They do not need any light whatsoever and can’t be blinded by direct sunlight.
There may be an issue with objects with no heat source being left on a crossing, for example an un-braked trailer. However, thermal imaging can usually even pick tis type of object out due to temperature differentials, so it may be a solution when combined with other detection technology.
Millimetre-wavelength beam interruption
Beam interruption is a method of obstacle detection based on microwaves. The transmitting antenna emits a beam signal to a transceiver. If an object enters the path of the beam, the signal is weakened to the transceiver, indicating its presence. With the use of reflectors and amplifiers it is possible to produce a weave of beams across an area. So, a beam could be placed across each entrance and diagonally from corner to corner to cover the crossing.
One of the world’s first Safety Integrity Level (SIL) 4 systems was installed in Italy and was based on millimetre wavelength beam interruption. While safe, the system was very sensitive to changes in temperature, rain and condensation on the transmitter and receiver sensors, which resulted in low availability. There was a need for periodic calibrations and maintenance and this was compounded by the need for a large number of sensors for the coverage of the whole area of the level crossing. The narrow beam width and limited field of view resulted in a requirement for even more sensors for high objects.
LiDAR covers the crossing area with pulses of near-infrared light that are reflected off the surface of an object on the crossing. The reflected pulses can then be analysed to determine its location, direction and speed.
Light has shorter wavelengths than radio waves, which means that LiDAR has the potential for more accuracy than radar. Network Rail used LiDAR in its first generation of OD crossings to supplement the radar to improve the detection of low objects. However, its improved sensitivity also means that it has the potential to be susceptible to small-sized objects, such as water vapour droplets that make up fog, although this can be mitigated by software algorithms.
Because LiDAR needs light to operate, the equipment must be located in a transparent housing and so is susceptible to water, dirt and dust on the glass. This requires maintenance and can result in low availability. LiDAR has a narrow beam width and a limited field of view, therefore additional sensors may be required for high and low obstacles. Some systems have mechanical moving parts which can result in a low mean time between failures (MTBF).
An induction loop consists of a cable, containing a coil wire transceiver (transmitter and receiver), arranged in a loop in order to create an electromagnetic field. It is used to detect metal objects, so is not of any use for pedestrian detection which would require another technology. The loop emits an electromagnetic field and a metallic object entering the looped area disturbs the field and induces a current. The output of the loop is fed into a processor and analysis determines the speed and size of the object passed over the field.
Unfortunately, there are an increasing number of road vehicles using composite and aluminium materials, which provide less of an induced current than steel, and problems have been reported in detecting lorries with high axles/ground clearance. Another difficulty is installing and maintaining induction loops in the surface of the crossing deck or road.
A strain gauge can be used to measure the deformation (strain) of a material. A strain gauge could be installed in a crossing, which would detect deformation of the crossing decking when a vehicle travels over it. The strain gauge could also use fibre optic technology. A strain gauge should be able to be calibrated for both vehicles and pedestrians, but may not be able to detect small children.
A similar detection technology is piezometers, which are made from rugged, weatherproof semi-conductor materials and could be laid in the crossing decking. Deformation of the piezometer caused by the weight of an object changes the conductivity and this can be analysed by a detector to identify the presence of objects.
Piezometers and strain gauges have the potential of being more reliable than induction loops, but locating the detectors in the crossing decking makes them as difficult to maintain.
These are designed to detect the presence of objects by changes in the frequency of sound waves. The sensor emits ultrasonic sound pulses that can’t be heard by the human ear. When the pulse reaches an object, the sound is reflected by the surface.
At a level crossing, ultrasonic sensors would have to be suspended above the crossing area and be able to emit the sound waves onto the crossing decking. Multiple sensors would be required to avoid black spots and, on electrified routes, the equipment would be in very close proximity to parts of the overhead wires.
The equipment would be more susceptible to vandalism and damage from members of the public since it is much more prominent at a crossing than other forms of detection. However, it is reported that the USA has performed successful obstacle detection trials using an array of ultrasonic sensors suspended over a level crossing.
This uses radio waves to detect objects, which is why it’s called radar – radio detection and ranging. The radar detector transmits radio waves over an area and monitors for any echoes. If an echo is received, this indicates that a wave has hit a surface of an object and has been reflected back.
By analysing the echo, the distance, position and speed of an object can be determined. The distance of the object can be identified by the time taken for the echo to return to the source or, if the radar uses frequency modulation, the distance can be determined by the difference between the emitted frequency and the echo. The distance and direction can then be used together to determine where an object lies within the level crossing. Reflectors can be installed on the boundaries of the crossing to obtain a reference echo signal and use it to monitor the status of the area and of the sensor itself.
OD radar systems have been developed with SIL 4 integrity. These now include systems with a wide beam width, so there is no need for multiple sensors for high and low obstacles, and which are able to detect obstacles of any harmful material or persons (including children) of any orientation. Radar-based systems are able to reliably detect objects through rain, fog, snow, hail, and, with no mechanical moving parts, they are able to deliver high availability.
One benefit of radar over other means of detection is that some safe low-density material objects, for example an empty paper box, will be ignored. Being radio-based, a radar OD system will normally require a radio license, but this does mean that the railway infrastructure manager will have exclusive use of the frequency and be will be able to manage any interference.
IDS Ingegneria Dei Sistemi is a world leader in providing radar systems for a variety of applications and has extensive research and development facilities. In September 2016, in partnership with Intecs SpA through the Stars Railway Systems consortium, the company provided its first level crossing system in Northern Italy. In total, more than 100 systems are on order – 49 have been delivered to date with all of them scheduled to be installed and operational by the end of 2018.
The system is able to monitor level crossings of any shape with, normally, only one sensor required per level crossing. However, up to four radar sensors can be used to cover all geometries. The sensor is able to operate in all weather conditions and has a predicted MTBF greater than 10 years. The system is capable of being configured to detect objects within 100mm to 200mm of a boundary, so there is a very low risk of objects or persons (vertical or horizontal) being missed. All current installations are performing as designed, with no false alarms reported.
In the UK, IDS leads the provision of specialist electromagnetic modelling software and consultancy services to the Ministry of Defence and other customers in the defence industry, so it has access to all the skills and knowledge required to provide radar systems for all rail applications.
Thanks to Paul Davies of IDS Ingegneria Dei Sistemi for his assistance with this article.
This article was written by Paul Darlington.