Technology of ATP/ATC systems
This chapter illustrates the principles of the design of the main classes of ATP system, without claiming to be exhaustive. Although we begin our journey with the oldest and simplest systems and their technology may appear to obsolete, many of these are still in existence today. In part this is due to the fact that any ATP implementation requires substantial investments in both infrastructure and the fleet. Once selected, any given implementation of an ATP/ATC systems tends to lock the railway into the associated operational principles, rules and regulations. On top of that, technology choices for ATP/ATC, such as a type carrier and modulation for the track to train transmission create a mutual dependency with the equipment, e.g. a coded track circuit or radio bearer service, a set of mutually dependent EMC/EMI limits etc. If not mandated by law, the business case for ATP/ATC is difficult to make, as shown in the UK with the ERTMS/ETCS business case and the business case for migration from a given system, generation or technology to another is even worse.
In the field of course, many permutations of the implementations described exist.
Intermittent Mechanical Systems
Early train stop systems were based on mechanical coupling between track and train. In its simplest form, the mechanical arm or bar of the trackside system moves up and down according to the signal aspect. When the signal is at danger and the train passes it, this arm hits an actuator of a brake valve (trip cock) and releases the air pressure of the brake system. Immediately a full brake application is initialised.
When signals changed from mechanical signals to colour light signals, the train stop arm on the trackside had to be moved either by magnets, electro-hydraulically or by motors. Similar systems are still in use on London Underground and the S-Bahn in Berlin.
The obvious disadvantages of such mechanical systems are the associated regular maintenance required for the mechanical parts and the “reset procedure” after a trip cock has been tripped, which not only is time consuming, but requires the driver to leave his cab and walk along the track which immediately exposes him to significant health and safety risks, not the least of which is being hit by another train.
Whilst such systems were generally regarded as fail-safe by the inherent mechanical properties of the design and equipment, reports of recent experience on LUL appears to has disproved this.
Intermittent Contactless Systems
Permanent Magnet Systems
The next evolutionary step for ATP systems was to replace the mechanical interaction between track and train by a “transmission system” based on magnetic coupling. A permanent magnet operates an on-board relay when the train is passing, which again opens a brake valve so that the brake is activated. If the signal is not at stop, a coil around the permanent magnet is powered and cancels the magnetic flux of the permanent magnet. When the train passes such a signal, there is no effect on the brake system. The system already shows a basic safety principle of railway signalling, a design based on the high probability of a reaction to the safe side in the case of a defect or malfunction we now call fail safety. In case of any failure of the active electrical part, the train would be stopped.
A similar principle is used in the British Automatic Warning System AWS, here a combination of different polarisation of magnetic fields of a permanent magnet and an electromagnet is used to either warn the driver, or simply to indicate that a magnet has been passed. In case of a warning, the driver has to acknowledge the warning within four seconds.
An electrical transmission by a physical contact between a trackside contact rail and a trainborne brush is used in the French and Belgian “crocodile” system. Depending on the polarity of the received voltage on board, the driver is also either warned or informed.
AC intermittent systems are usually based on the principle of installing permanent magnets in the track, the field of which can be cancelled by an AC signal which is fed through additional coils when the signal shows a proceed aspect. Examples of such systems include the German Indusi system, the Glasgow Subway  CTS system etc.
The German INDUSI system uses tuned circuits trackside and on board. The circuits are tuned to the same frequency and are detuned when a close coupling between both circuits takes place. The on-board resonant circuit is a serial resonant circuit and is fed by an oscillator with its natural resonant frequency. A high oscillator current is therefore obtained. The oscillator current is monitored by the train borne equipment of INDUSI. When the train passes a trackside resonant circuit (still referred to as a magnet!), the coupling detunes the train borne equipment and the oscillator current decreases. The oscillator current is compared to a threshold and when the current falls under this threshold, a reaction is triggered. To cancel the operation of the trackside resonant circuit in case of a go-aspect, the capacitor on the trackside circuit is shorted by the signalling equipment and the train resonant circuit is only slightly damped when passing the signal, the threshold would not be touched.
In the INDUSI system, 500 hertz, 1000 hertz and 2000 hertz circuits are used. The 2000 hertz magnet is placed at a main signal, the 1000 hertz magnet is at a distant signal and the 500 hertz magnet at a distance of 450 meters before the point of danger, which in general is at the end of the overlap.
In the traditional INDUSI system, supervision was based on speed supervision when passing the magnets or fixed times after that occurrence. In the newer train born equipment which is based on micro-processor technology, a continuous supervision of braking curves which is triggered when passing a magnet is performed by the software logic.
Transmission Based Intermittent Systems
A common feature of the intermittent systems described previously is their rather limited “data transmission capacity”, usually equivalent to 1 bit of information (signal on or off).
Modern intermittent systems are based on transmission of data-telegrams to trains. These telegrams are always secured by a safety code, so that sporadic interference cannot cause an undetected corruption of the message. Existing systems are either based on transponders with a contact length for data transmission of less than 1 meter or on loops which can have lengths of up to a few hundred meters.
A typical representative of the transponder solution is the KVB system used by SNCF, which in turn was derived from the Swedish Ebicab system. A minimum of two transponders is laid out in the middle between the two rails. The duplication of the transponders primarily guarantees the safety of the system and also enables a logical detection of travel direction. The locomotive is emitting pulses of 27 MHz carrier frequency at a repetition rate of 50 kHz. The trackside transponder transforms these pulses to a frequency of 4.5 MHz and reflects them to the locomotive. The shape of the amplitude of the reflected signal indicates if a logical one or a logical zero is transmitted. This system needs no trackside power supply for the transmission itself and hence no cabling if the information in the transponder or balise is fixed. Active transponders and/or transponders transmitting variable data obviously do need to be cabled.
A further development of the transmission principle of Ebicab and KVB is applied in the ERTMS/ETCS Eurobalise.
Transponder systems can be active or passive, the latter employing remote powering of the transponder by the vehicle. Transponders can also be fixed in their message content, or contain switcheable messages that can be selected through a connection to a signal, interlocking etc. Transponders that are both passive and fixed have the advantage of requiring no trackside cabling at all. Selectivity of a balise’s message to the direction of travel of a train can be achieved physically, e.g. by offsetting them from the centre of the rails, or through coding of messages in two or more balises in a group etc. The latter is also used to extend the amount of data that can be transmitted. Balise messages can contain pointers, identifying the next balise a train should encounter, depending on its route if needed, to ensure an alarm can be raised if the next balise is missed, misread or simply absent.
Loop systems always are active, that means that they require a trackside power feed.
The system installed in the Netherlands as ATBNG, Belgium (TBL) and the UK on the GWML is the TBL system, which uses a combination of an active transponders and loops. Both parts are fed with the identical information, but at different transmission speeds. The transponder transmits at a bit rate of 25 kbit per second, the loop with approximately 1 kbit per second. Transponder and loop are transmitting at a carrier frequency of 100 kHz. In this system transponders are located at signals, whereas the loops are used to transmit semi-continuous “infill” information in rear of a signal where needed. The function of the infill information is to allow a speed supervision curve to be updated as soon as a signal aspect improves, rather than forcing the train to continue braking until it reaches the transponders at the location of the next signal.
The SELCAB system is based on loop transmission only. It will be described in more detail later on.
Future intermittent systems could possibly be based on microwave transmission. One example is the AMTECH system, which was standardised by UIC for wagon identification application (AVI). This system could be used in a reverse manner, that means the transponder is mounted on trackside and the reader on trainborne for an intermittent data transmission system. This AMTECH transponder is based on the backscatter principle and is operating at a frequency of 2.4 to 2.5 GHz. The transmission speed is in the range of 150 kbit per second. The microwave technology equipment is small in size and low in cost, but is sensitive to transmission problems when debris, salt water etc. is present in the transmission path. Although it was considered to be a promising technology especially where fixed information has to be transmitted, these transmission problems prevented its selection as the transmission medium for Eurobalise. At present, the application of this system as transponder is limited to Metros (London and Madrid).
The traditional method for data transmission in continuous systems is to use coded track circuits. In the first incarnations of this family of ATP systems the conventional low frequency track circuits used throughout the world are used as the carrier for track to train transmission. The detection current flowing through the rails as soon as a section is occupied, is used as the carrier for the ATP/ATC codes and is “amplitude modulated” by periodically switching it on and off with a frequency of a few Hertz, indicating the speed code. The carrier frequency itself can be 50 or 60 Hz depending on the mains frequency of the country, or 75 Hz as in the Dutch first generation ATB system, chosen to avoid interference from the public grid. The encoding of the speed codes on trackside and decoding on the trainborne reception end was traditionally done means of tuned mechanical relays. The supervised speed depends on the speed code, i.e. the modulation frequency at which the track circuit current is switched on and off. These devices are simple, robust and used to be cost effective, as they have very low numbers of parts. There are also certain disadvantages for this type of continuous ATP/ATC system namely:
• The train with the worst braking characteristics defines the braking profile, faster trains with better brakes have to brake early.
• The brake supervision is very coarse, due to the small amount of sped codes available, and usually a long overlap has to be provided (typically one block section).
• The train always has to move from receiver to transmitter of the track circuit, bi-directional running requires a switchover of transmitter and receiver.
• High return currents and / or harmonics etc. generated by modern traction systems and all sorts of disruption in the “code transmission, e.g. at block joints, special trackwork etc. can cause corruption of the information as these systems do not employ any for of data protection or redundancy.
In more recent systems, the same principle is used, but upgraded to increase the amount of information transmitted. In the Italian BACC system, a second carrier with a higher frequency is used to transmit additional information for high speed trains.
The French TVM 300 and 430 systems use jointless track circuits and more advanced frequency modulation to transmit the information from track to train.
All continuous systems described above transmit permissible maximum speeds to the train. The braking curve is based on one or two typical train types. These systems were developed for lines with one or two types of trains only. Mixed traffic always results in a non-optimal use of the line capacity.
The German LZB system uses bi-directional data transmission through cable loops laid out between the rails. The system was specified by UIC and operates at a carrier frequency of 36 kHz to the trains and 56 kHz from the trains to the trackside equipment. The data speeds are 1200 bit per second and 600 bit per second respectively. This system is mainly based on transmission of target values rather than speeds. Therefore, any mixture of train classes can be handled by this system. The individual train characteristics are known on board and the speed calculation is done on board based on the target values from trackside and the train characteristics. A more detailed description of the LZB system is contained in chapter 0.
In the 1980s and 1990s a number of research programs such as ASTREE and DIBMOF investigated the use of continuous systems be based on radio transmission. The experience gained was used in the ERTMS projects following the decision by the participating railways to select GSM-R as their future standard for all train radio applications. The specifications developed in the Eirene project have subsequently been mandated through EU legislation as the European standard.
In radio based systems, transponders containing i.a. position reference data have to be provided to support the on board odometry. In rail based continuous transmission ATP/ATC systems this functionality was not needed, or deduced from such things as the sectioning of the track circuits or crossings of the loop cable. In a moving block system, the train positioning has to be especially accurate. A combination of various sensors such as radar, tachogenerator and accelerometer has to be used and/or positioning transponders have to be laid out in rather short distances. The physical separation of transmission channels of both tracks of a line also has to be replaced by information through transponders. This requires a high number of transponders, which again requires cost effective solutions and favours the passive fixed transponder solution.