GB2597348A

GB2597348A - Monitoring trains on a railway

Info

Publication number: GB2597348A
Application number: GB2103462.4A
Authority: GB
Inventors: Wing Chan Ho; Pan Tam Wai; Chun Ng Ka; Lun So Hon; Kin Ng Wai
Original assignee: MTR Corp Ltd
Current assignee: MTR Corp Ltd
Priority date: 2020-03-26
Filing date: 2021-03-12
Publication date: 2022-01-26
Also published as: GB202103462D0

Abstract

Trains are monitored by receiving data 218, 228 relating to positions and speeds of multiple trains; identifying first train 210 as a collision risk candidate of second train 220; using their positions or speeds to determine whether collision risk criteria 224A are met and, in response, causing the second train to activate an event 224B (e.g. driver alert or automatic braking). The candidate may be identified by a system 224 on-board the second train or by a remote server 230. It may be identified based on both trains being on the same railway section between two turnouts or within a zone around a rail-rail junction. Alternatively, trains are monitored by receiving their positions and speeds, identifying a nearest leading train in front of a following train and activating an event in the following train if a difference between a braking distance of the following train and a sum of a separation distance and braking distance of the lead train passes a threshold.

Description

MONITORING TRAINS ON A RAILWAY

FIELD

The present disclosure relates to a method, system and apparatus for monitoring trains on a railway.

BACKGROUND

Rail transport is a mode of transport in which the vehicles, known as trains, are guided by one or more tracks known as railway tracks. A railway may include one or more railway tracks linking a number of stops at which passengers may board or disembark, or at which goods may be loaded or unloaded.

Light rail transit (LRT) is a form of urban railway which serves urban areas and has stops which are relatively close together. LRT railways may run in close proximity to urban roads and may have many road-rail junctions at which a road crosses the railway or crossings at which pedestrians may walk over the railway. Compared to conventional railways, LRT railways have more frequent stops and a much larger number of junctions and crossings. Therefore it is necessary for the driver to take into account the presence and movement of road vehicles and pedestrians and for instance to look out for road traffic signs.

Due to the close proximity to pedestrians and road vehicles, which may cross the railway tracks, light rail transit systems are often operated in a manual driving mode, which relies on the driver to control the speed of the train and apply brakes when necessary. Human errors such as negligence, attention lapses, poor physical conditions or misjudgement by the train driver may cause safety or other service related issues.

SUMMARY

A first aspect of the present invention provides a method of monitoring trains on a railway comprising: a) receiving data relating to positions and speeds of a plurality of trains; b) identifying a first train of the plurality of trains as a collision risk candidate in respect of a second train of the plurality of trains; c) determining whether predetermined collision risk criteria are met based on the position and/or speed of the first train and the position and/or speed of the second train; and d) causing the second train to activate an event in response to determining that the predetermined collision risk criteria are met.

A second aspect of the present disclosure provides a method of monitoring railway trains on a railway comprising: receiving positions and speeds of a plurality of trains; identifying a first train of the plurality of trains as a leading train and a second train of the plurality of trains as a following train, wherein the leading train is a nearest train in front of the following train along a route of the following train and the leading train and following train are travelling in a same direction; determining a current separation distance (DS) between the first train and the second train; determining a braking distance (DL) of the first train and a braking distance (DF) of the second train; and causing an event to be activated on the second train in response to a difference between the braking distance of the second train and the sum of the current separation distance and the braking distance of the first train (Ds-FDLDF) passing a predetermined threshold.

A third aspect of the present disclosure provides a server comprising a processor and a machine readable storage medium storing instructions executable by the processor to: receive data relating to positions and speeds of a plurality of trains from on-board systems of the plurality of trains; identify a first train of the plurality of trains as a collision risk candidate in respect of a second train of the plurality of trains; and notify an on-board system of the second train that the first train is a collision risk candidate in respect of the second train.

A fourth aspect of the present disclosure provides an on-board system of a train comprising: a GPS receiver to generate data relating to a position of the train; an odometer to generate data relating to a speed of the train; and a controller configured to: transmit data relating to the position and speed of the train; receive data relating to the position and speed of another train; determine whether predetermined collision risk criteria are met based on the position and/or speed of the train and the position and/or speed of the another train; and cause the train to activate an event in response to determining that the predetermined collision risk criteria are met.

A fifth aspect of the present disclosure provides a system for monitoring trains comprising one or more apparatus which are configured to perform the method of the first or second aspects of

the present disclosure.

The system may comprise a server according to the third aspect. The system may comprise one or more trains comprising on-board systems according to the fourth aspect.

Further aspects and features of the present disclosure are provided in the description and claims appended hereto.

BRIEF DESCRIPTION OF THE DRAWINGS

Examples of the present disclosure will be explained below with reference to the accompanying drawings, in which:-Fig. 1 is a schematic diagram of a method of monitoring trains according to an example of the present disclosure; Fig. 2A is a schematic diagram of a system for monitoring trains according to an example of the

present disclosure;

Fig. 2B is a schematic system of an apparatus for monitoring trains according to an example of the present disclosure; Fig. 3 is a schematic diagram of an on-board system of a train according to an example of the

present disclosure;

Fig. 4A shows an example of a first train and a second train on a same railway track and travelling in a same direction according to an example of the present disclosure; Fig. 4B shows an example of a plurality of trains on a railway at a first point in time according to an example of the present disclosure; Fig. 4C shows an example of a plurality of trains on a railway at a second point in time

according to an example of the present disclosure;

Fig. 4D shows an example of a plurality of train routes on a portion of a railway including a plurality of rail-rail junctions according to an example of the present disclosure; Fig. 4E shows an example of a method of identifying a leading train according to an example of

the present disclosure;

Fig. 5 shows an example method of monitoring trains according to the present disclosure; Fig. 6 shows an example method of monitoring trains according to the present disclosure; Fig. 7A shows an example of a pair of non-conflicting train routes in proximity to a junction according to the present disclosure; Fig. 7B shows an example of a pair of conflicting train routes in proximity to a junction according to the present disclosure; Fig. 8A shows an example of a first predetermined zone around a rail-rail junction according to the present disclosure; Fig. 8B shows an example of a second predetermined zone around a rail-rail junction according to the present disclosure; Fig. 8C shows an example method of activating an event in response to collision risk criteria being met in accordance with an example of the present disclosure; Fig. 9 shows an example method of activating an event in response to collision risk criteria for junction collision being met in accordance with an example of the present disclosure; Fig. 10 shows an example system for monitoring a train including a server and on-board system

according to an example of the present disclosure.

DETAILED DESCRIPTION

Various examples of the disclosure are discussed below. While specific implementations are discussed, it should be understood that this is done for illustrative purposes and variations with other components and configurations may be used without departing from the scope of the

disclosure as defined by appended claims.

In the context of this disclosure the term railway refers to one or more rail tracks for guiding the movement of trains which are to travel above the rails. For example, a railway may include a pair of parallel rails which interface with train wheels and guide the movement of the train. The term "train" means any vehicle which is designed to travel along the tracks of a railway including, but not limited to, heavy rail and light rail (LRT) trains. LRT trains are sometimes referred to as light rail vehicles (LRV).

The teachings of the present disclosure may find particular application where trains are driven in manual driving mode. This is more common on LRT, as the LRV trains often need to stop or moderate their speed to accommodate the movement of pedestrians and/or road vehicles. In contrast, while heavy rail trains may be driven in manual mode, they are more often controlled automatically by a block signalling system. In a block signalling system the railway is divided into blocks, an electrical signalling system associated with the rails detects on which block(s) of track a train is present and only one train is allowed to be present on each block at any one time.

It is envisaged that the teachings of the present disclosure will be primarily applied to light rail (LRT) railway systems, but may also be applied to other types of railway.

In manual driving mode it is helpful to inform the driver of a position and speed of the train. One way of determining the position and speed of the train is by use of a Global Positioning System (GPS) receiver and an odometer on the train. The GPS receiver may determine the position of the train, while the odometer may determine a speed of the train. For instance the odometer may determine the speed by measuring a speed of rotation of one or more wheels of the train.

This approach may provide more accurate position and or speed of the train, compared to a block signalling system which is of limited resolution.

Further the data from the GPS receiver and odometer may be combined to give more accurate estimations of speed and position than the readings of the GPS receiver or odometer alone. In some examples the train may also include a reader (such as a radio frequency identifier (RFID) reader) which is to read data sent by trackside transmitters (such as a RFID transmitter). The data sent by the trackside transmitters may for instance relate to a speed limit for that section of the railway. By combining the GPS, odometer and in some cases also RFID data, the driver of the train may be alerted if they are travelling above a speed limit for a particular section of the railway. Such approaches are discussed in MTR's patent HK1223787, which is incorporated herein by reference.

A plurality of trains may run on a railway between stops of the railway. If the railway has a number of lines forming a railway network, then there may be rail-rail junctions (hereinafter simply referred to as "junctions") between different lines on the railway. As the headway (time between successive trains) urban railways, especially but not limited to LRT railways, can be quite short, there is a risk of collision between trains running on the same length of railway track or crossing paths at a junction.

Fig. 1 shows an example method 1 for monitoring trains on a railway according to the present 30 disclosure.

At step 10 data relating to positions and speeds of a plurality of trains is received.

At step 20 a first train of the plurality of trains is identified as a collision risk candidate in respect of a second train of the plurality of trains.

At step 30 it is determined if the positions and/or speeds of the first train and the second train meet predetermined collision risk criteria.

At step 40, in response to determining that the collision risk criteria are met in step 30, an event is caused to be activated on the second train.

The event which is activated when the collision risk criteria are met may, for example, be an alert to a driver of the train (e.g. by visual alert on a driver control panel or an audio alert in the driver cabin), automatically activating a braking system of the train, or automatically reducing a speed of the train. In this context "automatically" means without the intervention of the train driver.

Thus it will be appreciated from the above that a first train is identified as a collision risk candidate in respect of a second train and an event is activated on the second train when certain predetermined collision risk criteria are met. This may help to prevent an accident by alerting the driver of the second train or automatically braking or reducing the speed of the second train when the collision risk criteria are met.

The steps of method 1 of Fig. 1 may be carried out by a server remote from the first train and the second train, by an on-board system of the first train and/or second train or a combination of both. An on-board system of a train is a system which is installed in, or carried on, the train itself and thus moves along with the train.

Fig. 2A shows an example of a system 200 for monitoring trains in which method steps 10 and 20 of Fig, 1 are performed by a server which is remote from the first train and the second train and method steps 30 and 40 are performed by an on-board system of the second train. This may be referred to as a distributed approach. The system 200 includes a first train 210, a second train 220 and a remote server 230 for monitoring the trains. The first train 210 includes an on-board system 214 for monitoring the speed and tracking a position of the first train and the second train 220 includes an on-board system 224 for monitoring the speed and tracking a position of the second train. The on-board system 214, 224 may be communicatively coupled to a radio frequency device for wirelessly communicating with the remote server 230.

The server 230 may for example be a physical server, a virtual server or a cloud computing service. The server 230 may include a module 240 for receiving data relating to positions and speeds of a plurality of trains on the railway and a module 250 for identifying collision risk candidates. Thus, module 240 may perform method step 10 of Fig. 1 and module 250 may perform method step 20 of Fig. 1. The modules 240, 250 may be implemented by way of machine readable instructions stored on a non-transitory storage medium of the server 230 and executable by a processor of the server.

As shown in Fig. 2A data 218 relating to the position and speed of the first train is received by the server 230 from an on-board system 214 of the first train. Likewise data 228 relating to the position and speed of the second train are received by the server 230 from an on-board system 224 of the second train. For example the data 218, 228 may be transmitted to the server 230 over a wireless network, such as a telecommunication network, by radio frequency devices 216, 226 of the first and second trains. The data may be received by the server 230 directly from the on-board systems, as shown in Fig. 2A, or received indirectly from the on-board systems via one or more intermediate systems or devices.

In the example of Fig. 2A the method steps 30 and 40 of Fig. 1 are implemented on the on-board system of a train. Thus the on-board system 224 of the second train includes a module 224A for determining if the positions and/or speeds of the first train and the second train meet predetermined collision risk criteria (step 30 of Fig. 1) and a module 224B for activating an event in response to the predetermined collision risk criteria being met (step 40 of Fig. 1).

This approach, in which method steps 30 and 40 are implemented by the on-board system of a train may allow the event to be activated in a timely fashion, even if there is a delay in communication between the server and the train, as the on-board system of the train can determine whether to activate the event itself, rather than waiting for an instruction from the remote server Thus, in the example of Fig. 2A, the server 230 receives speed and position data of a plurality of trains and identifies collision risk candidates for some or all of the trains based on this data, but the on-board systems of the trains determine whether or not there is an actual risk of collision based on the collision risk criteria.

In one example, the server 230 may determine that the first train 210 is a collision risk candidate for the second train 220 and may send a message 219 to the on-board system 224 of the second train indicating that the first train is a collision risk candidate. The message 219 may include the position and speed of the first train. The on-board system of the second train then determines whether collision risk criteria are met based on the speeds and/or positions of the first train and second train and activates an event on the second train if the collision risk criteria are met.

The server 230 may periodically send messages 219 to the second train so as to provide the on-board system of the second train with an up-to-date position and speed of the first train. Even if there is a delay in the wireless communication between the server and the on-board system of the second train, the on-board system of the second train may still ascertain its own current position and may determine the collision risk criteria based on the previously received position of the first train. As the trains only travel in one direction on the railway track, the current position of the second train will be closer to the previously received position of the first train than the actual position of the first train (which will be further ahead). Accordingly safety may be preserved even if there is a delay in receiving the position data of the first train.

While in the example of Fig 2A a remote server is used to perform steps 10 and 20 of Fig 1 and the on-board system of a train is used to perform steps 30 and 40 (a "distributed approach"), in other examples all of steps 10-40 may be performed by a remote server (a "server centric approach"). In that case step 40 involves sending a message to the on-board system of the second train to cause the event to be activated on the second train. However, if there are delays in wireless communication, then such a server centric approach may cause safety issues.

Another approach is for steps 10-40 to be performed by the on-board system of one of the trains, rather than a remote server. This may be thought of as a "peer to peer approach". In that case the monitoring of trains may be carried out by communicating data between on-board systems of trains rather than to a remote server that is separate from the trains.

For example, the on-board system of the first train may be configured to receive the data relating to the position and speed of the second train from an on-board system of the second train and may receive data relating to the speed and position of the first train from devices installed on the first train, such as a GPS receiver, odometer or RFID reader etc. Likewise, the on-board system of the second train may be configured to receive the data relating to the position and speed of the first train from an on-board system of the first train and may receive data relating to the speed and position of the second train from devices installed on the second train, such as a GPS receiver, odometer or RFID reader etc. In the peer to peer approach, the on-board system of a train may receive speed and position data from one or more trains. For example, the on-board system of each train may broadcast the speed and position data of the train and may scan for broadcasts from other trains. In this way, the on-board system of a train may be alerted to nearby trains and may identify some or all of the nearby trains as collision risk candidates as described above. In some examples, the onboard system of the train may receive speed and position data from nearby trains, but not from all trains on the railway. This simplifies the identification of collision risk candidates as there are less trains to consider For example, if the broadcast is by wifi or Bluetooth then the transmission of speed and position data may be directly from one train to another and may have a range limited by the line of sight and/or range of the wifi or Bluetooth protocol. For example wifi may have a range of a hundred meters or so. In this way each train may receive broadcast speed and position data of nearby trains, e.g. within 100m, without receiving such data from further away trains. In other examples, if the broadcast is by a telecom network such as 3G or 4G etc, then the broadcast may be to all trains on the network. However, the on-board systems of trains receiving the broadcast may filter based on position to exclude trains which are not nearby. In this context nearby may for example be considered to be a same or adjacent section of track or within a predetermined distance such as 100m. In some examples the train-to-train communication of speed and position data may use Vehicle-to-Everything (V2X) communication, for instance as defined by the 3rd Generation Partnership Project (3GPP).

Fig. 3 shows an example of an on-board system 300 for monitoring the speed and tracking the position of a train. The position and speed of the first train may be determined based on data from a GPS receiver and an odometer of the first train. Thus in the example of Fig. 3 the onboard system includes a GPS receiver 310, an odometer 320 and a controller 330 which is to determine a position of the train and a speed of the train based on data from the GPS receiver and the odometer. A plurality of trains running on the railway, including the first train and the second train, may carry an on-board system 300 as shown in Fig. 3.

If a distributed approach is used in which steps 10-20 are carried out by a remote server and steps 30-40 are carried out by the on-board system, then controller 330 of the on-board system may include modules for receiving the speed and position data and identifying the collision risk candidates, e.g. as shown by reference numerals 240 and 250 in Fig. 2A. These modules may be implemented as machine readable instructions stored on a non-transitory storage of the controller and executable by a processor.

If a peer to peer approach is used then the controller 330 of the on-board system may include modules for carrying out each of steps 10-40. If a server centric approach is used then the server may include modules for carrying out each of steps 10-40. The modules may be implemented as machine readable instructions stored on a non-transitory storage medium and executable by a processor.

An example of an apparatus which may correspond to a server of a server centric approach, or a controller of the on-board system of a train in a peer to peer approach, is shown in Fig. 2B.

The apparatus includes a processor 110 which may, for example, be a central processing unit, microprocessor or other type of processor. The storage medium 120 may, for example, be a hard disk, solid state memory, read only memory (ROM), random access memory (RAM) etc. The storage medium 120 stores instructions which are executable by the processor 110. The instructions may include at least the following: instructions 130 to receive data relating to positions and speeds of a plurality of trains, instructions 140 to identify a first train of the plurality of trains as a collision risk candidate in respect of a second train of the plurality of trains, instructions 150 to determine if the positions and/or speeds of the first train and the second train meet predetermined collision risk criteria and instructions 160 to cause the second train to activate an event in response to determining that the collision risk criteria are met.

Thus instructions 130, 140, 150 and 160 stored in the apparatus of Fig. 2B may correspond to method steps 10, 20, 30 and 40 respectively of the method 1 of Fig 1. VVhere the apparatus is a remote server, the instructions 130 to receive to receive data relating to positions and speeds of a plurality of trains may involve the server receiving speed and position data wirelessly transmitted by the on-board systems of the trains. Where the apparatus is a controller of an on-board system of a train in a peer to peer set up, the instructions 130 may involve the controller may receiving data relating to the position and speed of the train from the odometer and GPS receiver of the on-board system of the train and receiving data relating to the position and speed of one more other trains from messages wirelessly sent by the on-board systems of the one or more other trains Examples of the identification of collision risk candidates and application of collision risk criteria will now be described.

As discussed above, a collision risk candidate is identified and then it is determined whether there is a risk of collision with the collision risk candidate based on collision risk criteria. In response to the collision risk criteria being met, an event is activated on a train. One type of collision is a junction collision which is between trains in proximity to a rail-rail junction between different rail tracks. Another type of collision is a tail-head collision, which is between trains running a same railway track. These different types of collision may have different collision risk candidates and different collision risk criteria. Tail-head collision, which is between trains running in the same direction on a same railway track, will now be discussed.

Running on a same railway track means that the first train and the second train are running along a same length of the railway track in the same direction with one ahead of the other. Fig. 4A shows an example of a first train and a second train running on a same railway track. It can be seen that the first train 410 and a second train 420 are travelling along a same length of railway track 400 in a same direction (i.e. the direction from right to left in Fig. 4A). As the first train 410 is leading in front and the second train 420 is following behind, the first train may be designated as the "leading train" and the second train may be designated as the "following train". If the following train is travelling faster than the leading train, there may be a risk of collision, depending on the separation distance between the trains.

Each train follows a route which is a predetermined journey along the tracks of the railway from a starting point to a destination. A route may include a number of stops between the starting point and the destination. The railway may include a plurality of railway tracks, which may cross, merge with or split from each other at rail-rail junctions. A route may traverse several different railway tracks. Different routes may merge with other routes for part of their length before splitting off.

At any one time, for each train on the railway, there will be a train in front which is referred to as a leading train. For any given train ('the present train'), the leading train is the nearest train which is travelling in the same direction as the present train and is currently on the upcoming route of the present train. The leading train may change over time as trains follow different routes along the railway. For example, with reference to Fig. 4B which shows a part of the railway 401 at a first point in time, it can be seen that the black train 422 is a following train and the white train 421 is a leading train. Put another way, with respect to the black train 422, the white train 421 can be identified as the leading train. Meanwhile the grey train 423 is travelling in an opposite direction to the black train 422. However, at a later time shown in Fig. 4C, it can be seen that the white train 421 has split off on a track to the left and is no longer in front of, nor on an upcoming route of, the black train 422. However, the grey train 423 is now on the same railway track as the black train 422 and ahead of the black train 422. Thus, at the later time shown in Fig. 4C, the grey train 423 is the leading train with respect to the black train 422.

There may be a large number of routes which merge with or cross over each other. This is particularly common for LRT railways. An example is shown in Fig. 4D, where it can be seen that northbound routes 612, 614, 615 and 751 start on rail track 502 at the bottom of the figure and go to the top of the figure, but route 751 splits off to the left on track 502a at turnout A. Meanwhile southbound routes 612, 614 and 615 start at top of the figure on track 503 and go to the bottom of the figure, but westbound route 761P starts at the top on track 503 and splits to track 503a at turnout B and joins with track 502a at turnout C. Meanwhile, northbound route 761P and southbound route 751 start at the top left of the figure on track 504, but northbound route 761P splits off at turnout D and joins track 502 at turnout E. Meanwhile, southbound route 751 merges with track 503 at turnout F Thus it will be appreciated that identifying the leading train with respect to a present train is not straight forward, especially where there are a large number of trains travelling on different routes. Further, the identity of the leading train will change over time depending on the other trains on the railway and the routes which the trains are following.

Fig. 4E shows an example method 430 of identifying a leading train (e.g. the first train 410 in Figs. 1 and 4A), which is a collision risk candidate with respect to a present train (e.g. the second train 420 in Figs. 1, 2A and 4A).

At block 431 the railway is divided into a plurality of sections. Each section of the railway is a length of railway track between two turnouts or between a turnout and a railway endpoint. A turnout is an installation which enables trains to be guided from one track to another track, for instance at locations where tracks merge or divide.

Dividing the railway into sections in block 431 may be carried when setting up the train monitoring apparatus. For example, the railway operator may designate each position on the railway as belonging to a particular section and store this information in a database. The information may be stored as a map of the railway on the remote server Cif a distributed or server centric approach is used) and/or on an on-board system of the train (if a peer to peer approach is used).

In one example the railway operator defines point identifiers (IDs) at regular intervals along the railway. For example point IDs may be defined every 2m along the tracks of the railway. The position of the train, as determined from the GPS receiver and/or the odometer of the train, may then be mapped to the nearest point ID, so that at any one time a train has a point ID. This mapping may be carried out by the remote server or by the on-board train system. Each section of the railway is between a pair of turnouts or between an endpoint (such as the end of a line) and a turnout. Thus each point ID may be associated with a particular section of the railway. One section of the railway may span a plurality of point IDs. For instance if the section is 20m long and point IDs are every 2m, then the section would include 10 point IDs. An operator of the railway may assign each point ID to a section of the railway and store the correlation in a database or lookup table.

At block 432 each train on the railway is mapped to a railway section based on a position of the train. This may be carried out periodically or in real time. In one example the train may be mapped to a railway section by determining a point ID of the train and looking up a section of the railway do which the point ID belongs. If the method uses a peer to peer approach, rather than a remote server to identify the leading train, then only trains which the on-board system is aware of (e.g. through receiving broadcasts) may be mapped to sections of the railway, rather than all trains being mapped.

At block 433 it is determined whether there is another train on the same section of railway as the present train. If there is another train (i.e. at least two trains) on the same section of railway, then the method proceeds to block 434 and if not, then the method proceeds to blocks 435 and 436.

At block 434, the leading train is identified based on the positions of the trains on the section of railway. The leading train will be the train which is the closest to the present train and in front of the present train. For example, this may be determined by comparing point IDs of the trains or absolute positions of the trains.

At block 435, subsequent sections of railway on the route of the present train are checked for presence of trains. A subsequent sections of railway are sections of railway on the route of the present train, which lie in the direction of travel of the present train and which the present train has not yet reached. For example, subsequent sections of railway may be examined in turn until finding a section on which a train is present.

At block 436, the train which is closest to the present train on subsequent sections of railway is identified as the leader train (and thus a collision risk candidate) in respect of the present train.

Thus, the method identifies a first train as a collision risk candidate for the second train (the 'present train') based on the first train being on a same section of the railway as the second train, wherein a section of the railway is a length of railway track between two turnouts or between a turnout and a railway endpoint. Where there is no other train on the same section of railway, the method identifies the first train as a collision risk candidate for the second train based on the first train being a closest train to the second train on subsequent section(s) of the railway on the route of the second train.

This method of finding a collision risk candidate for head-tail collision may be relatively computationally efficient, as only trains on the same section of railway (blocks 433 and 434 of Fig. 4E) or subsequent sections of railway (blocks 435 and 436 of Fig. 4E) are considered rather than all trains on the railway. Further, in some examples only subsequent sections up to the next section on which a train is present need be considered rather than all subsequent sections.

It will be appreciated that the sections of the railway are between turnouts and thus different to the blocks of a conventional railway signalling system, in which the railway is divided into blocks by joints between adjacent rails. As a result the sections may be longer than blocks. Further, in a conventional railway signalling system only one train may be present on a block at a time, whereas a plurality of trains may be present in a section of the railway according to the present disclosure. VVhereas the location of a train in a block signalling system may be determined from electrical readings from a track circuit which passes electricity through the track, the location of the train in the present disclosure is determined based on an on-train system which may utilise a GPS receiver and an odometer Fig. 5 shows an example method 500 of applying collision risk criteria for a pair of trains running on a same railway track, where one train (the 'first train') is a leading train and the other train (a 'second train') is a following train. The method may be implemented by an on-board system of a train or a remote server At block 510 a current separation distance between the first train and the second train is determined. For instance in Fig. 4A, the trains are separated by a distance ("the current separation distance") shown by arrow 415. The separation distance may, for example, be determined based on the received position of the first train and the received position of the second train.

At block 520 a safe separation distance between the first train and the second train is calculated.

At block 530 it is determined whether the current separation distance is less than the safe separation distance.

Blocks 510 to 530 thus determine whether the collision risk criteria are met. The collision risk criteria in this case may be expressed as a separation distance between the first train and the second train being less than a safe separation distance.

At block 540, in response to determining that the current separation distance is less than a safe separation distance (i.e. in response to determining that the collision risk criteria are met), an event is activated on the following train (i.e. the second train 420 in the example of Fig. 4A).

The safe separation distance may be calculated based on a braking distance of the first train and a braking distance of the second train. The braking distance of the first train may be calculated based on the speed of the first train and a maximum braking deceleration of the first train, e.g. emergency braking. The braking distance of the second train may be calculated based on the speed of the second train, a normal braking deceleration of the second train, e.g. service braking, and a predetermined driver reaction time of the second train. The maximum braking deceleration and the normal service braking deceleration may for example be set based on the technical specification of the train and/or rules for driving of trains set by the railway operator. The instructions stored on the machine readable storage medium of the apparatus 100, 230 may include instructions to calculate the braking distance of the first train and braking distance of the second train in accordance with the above calculations and the method of Fig. 5 may include carrying out these calculations.

The predetermined driver reaction time may be a period of time set by the manufacturer of the train, or an operator of the railway, based on an estimated reaction time for a driver. The estimated reaction time may be determined based on tests, historical data or reference tables for instance. The estimated reaction time may depend upon the particular type of train or railway.

In one example, the braking distance of the first and second trains may be determined according to the equations below: Equation 1

DF

Equation 2 Where: D, is the braking distance of the leading train (e.g. the first train 410 in Fig. 4) D, is the braking distance of the following train (e.g. the second train 420 in Fig. 4) vs is the speed of the leading train Tv, is the speed of following train as" is the maximum braking deceleration rate of the leading train (as it is deceleration, it is a negative value) a; is the normal service braking deceleration rate of the following train (as it is deceleration, it is a negative value) T, is the predetermined reaction time of the driver of the following train When a difference between the braking distance of the second train and the sum of the current separation distance and the braking distance of the first train passes (e.g. is less than) a predetermined threshold the event may be activated.

This may be expressed mathematically as activating an event when: D2 + < predetermined threshold Equation 3 Where Ds is the current separation distance between the first train and the second train, D" is the braking distance of the first train and DF is the braking distance of the second train. Putting the above together the collision risk criteria may be expressed as: z Zac, Zat < predetermined threshold Equation 4 In some examples the predetermined threshold may be at least 10 meters.

Fig. 6 shows an example method 600 of applying the above collision risk criteria. The method may be implemented by a remote server, an on-board system of a train or a combination of the two.

At block 610 positions and speeds of a plurality of trains are received.

At block 620 a first train is identified as a leading train and a second train as a following train. The leading train is thus a collision risk candidate for the following train.

For example, the plurality of trains may include the first train, a second train and a number of other trains and the first train may be identified as a leading train based on the first train being on a same section of the railway as the second train, or the first train being a closest train to the second train on subsequent section(s) of the railway on a route of the second train. In this context, a section of the railway is a length of railway track between two turnouts or between a turnout and a railway endpoint.

In one example, the leading train and following train may be identified by using the method of Fig. 4E.

At block 630 a current separation distance (Ds) between the first train and the second train is determined.

At block 640 a braking distance (DL) of the first train and a braking distance (DF) of the second train is determined.

At block 650 it is determined whether a difference between the braking distance of the second train and the sum of the current separation distance and the braking distance of the first train (Ds+DL-DF) passing a predetermined threshold.

At block 660 an event is caused to be activated on the second train in response to a difference between the braking distance of the second train and the sum of the current separation distance and the braking distance of the first train (Ds+DL-DF) passing a predetermined threshold.

The predetermined threshold is thus an additional safety margin In one example the predetermined threshold is 10m.

Machine readable instructions for carrying out the above method may be stored on one or more non-transitory machine readable storage mediums and executable by one or more processors.

The braking distance of the second train may include a distance calculated based on the speed of the second train and a predetermined reaction time of a driver of the second train.

The method of Fig. 6 may further include any of the features of the methods described above in relation to Figs. 1 to 5.

Collision risk criteria in relation to head-tail collision risk candidates, e.g. trains running a same railway track, have been described above. Collision risk criteria in relation to trains in proximity to a rail-rail junction will be described below. Such collision risk criteria may be referred to as "side-on collision risk criteria".

A rail-rail junction is a junction between different railway tracks. For example, a point at which railway lines merge, or a point at which a railway line divides into a plurality of railway lines, or a point at which a first railway line crosses a second railway line.

Fig. 7A shows an example of part of a railway 700 including a number of rail-rail junctions. A first railway line 701 is for trains heading in direction left to right of Fig. 7A, for instance train 710 travelling in the direction shown by arrow 711. Meanwhile a second railway line 702, parallel to the first railway line 701 is for trains heading in direction right to left in Fig. 7A, for instance train 720 travelling in the direction shown by arrow 712. A third railway line 703 branches off from first railway line 701 at junction 701a. Third railway line 703 crosses over second railway line 702 at junction 703A and heads towards the bottom of Fig. 7A. Meanwhile, a fourth railway line 704 branches off from the second railway line 702 at junction 702a and merges with the third railway line 703 at junction 703b. It can be seen that junctions 701a, 702a and 703b are turnouts at which a train may change from one track to another, while junction 703a is an inter-crossing at which the lines cross each other but trains do not change tracks.

Referring to the example of Fig. 7A, it can be seen that the first train 710 and the second train 720 are in proximity to the various junctions 701a, 702a, 703a, 703b. However, as the route of the first train as shown by arrow 711 and the route of the second train as shown by arrow 712 do not conflict or cross over, there is no collision risk.

However, in the example of Fig. 7B, which shows the same part of the railway as Fig. 7A and in which like reference numerals denote like elements, there is a risk of collision. In Fig. 7B the first train 710 has a route 713 which diverts from the second line 701 and follows the third railway line 703 to the bottom of the figure. As can be seen, this conflicts with the route 712 of the second train 720. In another scenario, the second train 720 may follow a route 714 which diverts from the second railway line 702 and follows the fourth railway line 714 which then merges with the third railway line 703 at junction 702b. As can be seen in Fig. 7B this alternative route 714 of the second train conflicts with the alternative route 713 of the first train.

Accordingly, from Figs. 7A and 7B, it will be appreciated that the side-on collision risk depends on the particular routes which the trains are following. However, checking the side-on collision risk for all trains in the railway based on their projected future paths would be computationally intensive.

In one approach according to the present disclosure, the first train and the second train may be identified as collision risk candidates based on the first train being within a first predetermined zone around a rail-rail junction and the second train being within said first predetermined zone of said rail-rail junction. For instance both the first train and the second train may be within a predetermined distance of the rail-rail junction.

An example is shown in Fig. 8A, which shows a part of a railway 800 having a plurality of junctions in a layout similar to Fig. 4B. A first predetermined zone 810 surrounds the junctions.

The first predetermined zone 810 may be considered as a 'search window', as the apparatus is configured to search for trains which are collision candidates within the zone 810.

Fig. 8B shows an example of the same part of the railway 800 with a second predetermined zone 820 surrounding the junctions. The second predetermined zone is zone in which trains may collide if trains on different tracks are in the predetermined zone at the same time.

Examples of possible spaces occupied by three trains Ti, T2, T3 at the junctions in the second predetermined zone 820 are shown in Fig. 8B and it can be seen that these spaces overlap. Therefore the second predetermined zone may be thought of as a collision zone. As can be seen in Fig. 8B, the second predetermined zone 820 is smaller than the first predetermined zone 810. Put another way the second predetermined zone 820 is inside the first predetermined zone 810.

The boundaries of the first predetermined zone 810 may be determined based on the maximum speed limit for trains on the railway tracks near the junctions and the braking distances of trains used on the railway. For example, the boundaries of the first predetermined zone 810 may be defined such that if a train travelling within the speed limit outside the first predetermined zone (the 'search window') brakes, then it will come to a stop before reaching the collision zone 820. Thus the distance between the boundary of the search window 810 and the junction which the search window is associated with may be at least equal to a braking distance of a train travelling at the speed limit. The collision zone 820 may have boundaries defined by the length of the longest trains used on the railway, so e.g. the distance between a junction and a boundary of the junction's collision zone may be equal to or greater than a length of the longest train used on the railway. The length of the longest train used on the railway may be specified by the railway operator when setting up the apparatus.

As noted above, just because a pair of trains are collision risk candidates by virtue of them being in proximity to the junctions, e.g. within the search window, this does not mean there is necessarily an actual risk of collision. That will depend on the collision risk criteria which may be applied to determine whether there is an actual collision risk for which corrective action, such as activating an event on one of the trains, should be taken. In one example the collision risk criteria includes a first projected future space occupied by the first train being within a predetermined safety distance of a second projected future space occupied by the second train.

The first projected future space may be based on the position, speed, route path and length of the first train and a second projected future space may be based on the position, speed, route path and length of the second train.

Fig. 8C shows an example method of applying collision risk criteria for collision at a junction. The method may be implemented by an on-board system of a train and/or a by a remote server At block 810 a projected future space occupied by the first train is determined. For example this may be determined based on the position, speed, route path and length of the first train.

At block 820 a projected future space occupied by the second train is determined. For example this may be determined based on the position, speed, route path and length of the second train.

At block 830 it is determined whether the first projected future space of the first train is within a predetermined safety distance of a second projected future space of the second train.

At block 840 an event is activated on the first rain or the second train in response to determining that the first projected future space of the first train is within a predetermined safety distance of a second projected future space of the second train.

The method of Fig. 80 may be computationally intensive and does not definitively determine which train the event should be activated on. Therefore a further example method of operation 900 of the apparatus 100, 230 in respect of collision risk criteria for a pair of trains in proximity to a junction, is illustrated in Fig. 9. The method of Fig. 9 may be less computationally intensive and may determine on which train the event should be activated.

At block 910 a first train and a second train are determined to be collision risk candidates based on the first train and the second train being within a first predetermined zone (the 'search window") around a rail-rail junction, as described above.

At block 915A it is determined whether a route of the first train intersects with a route of the second train. If the routes intersect then the method proceeds to block 920. However, if the routes do not intersect then there is no risk of collision and the method proceeds to block 915B, where the method ends without activating an event one of the trains. In this way computational power is saved, as blocks 920-960 are not proceeded with if there is no conflict between the route of the first train and the route of the second train.

At block 920 it is determined whether the first train is within a second predetermined zone ('the collision zone') of the rail-rail junction. If yes, then the method proceeds to block 925.

At block 925 it is determined whether a projected future space of the second train extends into the second predetermined zone ('the collision zone') of the rail-rail junction.

The projected future space of the second train may be calculated based on the current speed of the second train, the braking deceleration rate of the second train (e.g. the normal service braking deceleration rate) and a buffer time. The buffer time may include a driver reaction time and a safety buffer. In one example, then projected future space may be mathematically expressed as: Equation 5 Where, DFspAcE is the projected future space; t*,,, is the speed of the second train; az is the normal service braking deceleration rate of the second train (as it is deceleration, it is a negative value); and T, is a buffer time.

The buffer time may be the sum of a predetermined reaction time of the driver of the second train plus a safety buffer, e.g.: Buffer time = driver reaction time + safety buffer Equation 6 If the projected future space does not extend into the collision zone, this indicates that the risk of collision is low as the train is some time away from the collision zone. However, if the projected future space extends into the collision zone, this indicates there may be a higher risk of collision.

In one example, the safety buffer may be expressed as a period of time and may be set such that the safety buffer multiplied by the maximum speed of the train is at least 10m. That way if the projected future space just touches the collision zone and the driver of the second train is instructed to brake at the normal braking rate now, the train can be expected to come to a stop approximately 10m away from the collision zone. Thus the safety buffer provides a safety margin.

At block 930 as the first train is in the collision zone and a projected future space of the second train extends into the collision zone, an event is activated on the second train. For example the event may be an alert to the driver of the second train, or automatically activating the braking system or reducing the speed of the second train.

If the determination at block 920 is negative, then the method proceeds to block 940.

At block 940 it is determined whether the first train has a braking distance within the second predetermined zone (the "collision zone") of the rail-rail junction. If the answer is yes, then the method proceeds to block 945.

At block 945 it is determined whether a projected future space of the second train extends into the second predetermined zone ('the collision zone') of the rail-rail junction. The projected future space of the second train may be calculated as described above for block 935.

At block 950, as the first train will be unable to stop before entering the collision zone and a projected future space of the second train is in the collision zone, an event is activated on the second train. For example the event may be an alert to the driver of the second train, or automatically activating the braking system or reducing the speed of the second train.

If the determination at block 940 is negative, then the method proceeds to block 960.

At block 960 it is determined which train has the lowest priority to occupy the junction. The priority of each train may be assigned by the railway operator and may, for example, be stored on the on-board system of the trains and/or the remote server. For instance certain routes may be given priority over other routes. Thus the priority of a train may depend upon the train's route.

At block 965 it is determined whether the train having the lower priority has a projected future space within the second predetermined zone ('the collision zone') of the rail-rail junction. The projected future space of the lower priority train may be calculated as described above for block 935. If the projected future space of the lower priority train extends into the collision zone, then the method proceeds to block 970.

At block 970 the event is activated on the train having the lower priority.

For example, if the first train has a lower priority than the second train, then the event is activated on the first train. However, if the first train has a higher priority than the second train, then the event is activated on the second train.

It will be appreciated that blocks 910 and 915 of Fig. 9 identify collision risk candidates, blocks 920, 940 and 960 determine which train an event should be activated on if collision risk criteria are met and blocks 925, 945 and 965 determine whether the collision risk criteria are met. Meanwhile an event is activated at blocks 930, 950 and 970.

The method 900 of Fig. 9 may be implemented by an on-board system of a train (peer to peer approach), by a remote server (server centric approach) or a combination of a remote server and an on-board system of a train (distributed approach).

In one example, a distributed approach is used in which blocks 910, 915, 920, 940 and 960 are performed by a remote server, while blocks 925, 930, 945, 950, 965 and 970 are performed by an on-board system of a train.

If a peer to peer approach is used then blocks 910-970 may be implemented by the controller of an on-board system of a train. In that case the on-board system of the train may determine whether the train is within the search window or collision zone by comparing the position of the train with information relating to the locations of the search window or collision zone, which may be stored on a storage accessible to the controller. In another example, the on-board system may determine the train has entered a search window or collision zone by receiving a signal from a trackside device, such as a RFID transmitter, indicating that the train has entered a search window or collision zone. The on-board system of the train may determine whether another train is within a search window or collision zone based on information received from the on-board system of the another train.

In one example, a system for the monitoring trains on a railway may include a remote server as described above together with a first train including a first GPS receiver and a first odometer and to generate data relating to the position and speed of the first train and a second train including a second GPS receiver and a second odometer and to generate data relating to the position and speed of the second train.

It will be appreciated from the above that data relating to the position and speed is generated by an on-board system of a train and the data is sent to the remote server. Fig. 10 shows an example system 1000 for monitoring trains on a railway including a remote server 1100, an onboard system 1300 of a train and a track side system 1200.

The on-board system 1300 may include an odometer 1310, a GPS receiver 1320 and a controller 1330 as described previously in relation to Fig. 3. The controller 1330 is configured to receive position and speed data from the GPS receiver and the odometer and send the data to the remote server 1100 over a wireless network, such as a telecommunication network 1400. For example the controller 1330 may be connected to radio frequency circuitry 1340 for wirelessly transmitting the data. The speed and position data may be processed by the controller 1330 before transmitting to the remote server The on-board system 1300 may further comprise a vehicle system interface 1360 for sending control signals to a train control system. For example, the vehicle system interface 1360 may allow alerts to be sent to the driver of the train via a user interface, display and/or audio system of the train. The vehicle system interface 1360 may also allow control signals to be sent to automatically activate a braking system of the train or to slow down the train by automatically reducing the engine speed etc. The on-board system 1300 may further comprise a radio frequency ID (RFID) reader 1350 which may be configured to read RFID transmitters 1210, such as RFID tags placed at a side of the track. Such RFID transmitters 1210 may store and transmit data relating to a status of the track, such as speed limits and/or location etc. The RFID transmitters 1210 may together form a trackside RFID network. The data from the RFID transmitters may for example be used to activate speed related events in the train, such as alerting the driver if a speed limit is exceeded, or to validate or calibrate speed or position data generated by the odometer and the GPS receiver.

It is to be understood that any feature described in relation to any one example may be used alone, or in combination with other features described, and may also be used in combination with any features of any other of the examples, or any combination of any other of the examples.

Claims

CLAIMS1 A method of monitoring trains on a railway comprising: a) receiving data relating to positions and speeds of a plurality of trains; b) identifying a first train of the plurality of trains as a collision risk candidate in respect of a second train of the plurality of trains; c) determining whether predetermined collision risk criteria are met based on the position and/or speed of the first train and the position and/or speed of the second train; and d) causing the second train to activate an event in response to determining that the predetermined collision risk criteria are met.
2. The method of claim 1 wherein steps a) and b) are performed by a remote server and steps c) and d) are performed by an on-board system of the second train.
3. The method of claim 2 wherein the remote server sends a message to an on-board system of the second train informing the on-board system of the second train of the position and speed of the first train.
4. The method of claim 1 wherein wherein steps a) to d) are performed by an on-board system of the second train.
The method of claim 4 wherein each of the plurality of trains has a respective on-board system which determines a position and speed of the train and wirelessly broadcasts the position and speed of the train for reception by nearby trains.
6 The method of any one of the above claims wherein the position and speed of the first train are determined based at least on data from a GPS receiver and an odometer of the first train; the position and speed of the second train are determined based at least on data from a GPS receiver and an odometer of the second train.
7 The method of any one of the above claims wherein the event is at least one of an alert to a driver of the second train and causing the second train to reduce speed or activating a braking system of the second train automatically.
8. The method of any one of the above claims wherein the first train is identified as a collision risk candidate for the second train based on the first train being on a same section of the railway as the second train, wherein a section of the railway is a length of railway track between two turnouts or between a turnout and a railway endpoint.
9 The method of any one of claims 1-7 wherein the first train is identified as a collision risk candidate for the second train based on the first train being a closest train to the second train on subsequent section(s) of the railway on the route of the second train.
10. The method of any one of the above claims wherein the collision risk criteria include a separation distance between the first train and the second train being less than a safe separation distance.
11. The method of claim 10 wherein the instructions include instructions to calculate the safe separation distance based on a braking distance of the first train and a braking distance of the second train.
12. The method of claim 11 wherein the instructions include instructions to calculate the braking distance of the first train based on the speed of the first train and a braking deceleration of the first train.
13. The method of claim 11 wherein the instructions include instructions to calculate a braking distance of the second train based on the speed of the second train, a braking deceleration of the second train and a predetermined driver reaction time of the second train.
14. The method of any one of claims 1-7 wherein the first train is identified as a collision risk candidate for the second train based on the first train being within a first predetermined zone around a rail-rail junction and the second train being within said first predetermined zone of said rail-rail junction.
15 The method of claim 14 wherein the collision risk criteria include a first projected future space occupied by the first train being within a predetermined safety distance of a second projected future space occupied by the second train, the first projected future space being based on the position, speed, route path and length of the first train and a second projected future space being based on the position, speed, route path and length of the second train.
16. The method of claim 14 wherein the collision risk criteria include the first train being in a second predetermined zone of the rail-rail junction.
17. The method of claim 14 wherein the collision risk criteria include the first train having a braking distance within the second predetermined zone of the rail-rail junction.
18. The method of claim 14 wherein the collision risk criteria include a route of the second train overlapping with a route of the first train.
19. The method of claim 14 wherein the collision risk criteria include a projected future space of the first train being within predetermined safety distance of a projected future space of the second train.
20. A method of monitoring railway trains on a railway comprising: receiving positions and speeds of a plurality of trains; identifying a first train of the plurality of trains as a leading train and a second train of the plurality of trains as a following train, wherein the leading train is a nearest train in front of the following train along a route of the following train and the leading train and following train are travelling in a same direction; determining a current separation distance (Ds) between the first train and the second train; determining a braking distance (DL) of the first train and a braking distance (DF) of the second train; and causing an event to be activated on the second train in response to a difference between the braking distance of the second train and the sum of the current separation distance and the braking distance of the first train (Ds+DL-DF) passing a predetermined threshold.
21 The method of claim 20 wherein the plurality of trains includes the first train, a second train and a number of other trains and wherein the first train is identified as a leading train based on the first train being on a same section of the railway as the second train or the first train being a closest train to the second train on subsequent section(s) of the railway on a route of the second train, wherein a section of the railway is a length of railway track between two turnouts or between a turnout and a railway endpoint.
22. The method of claim 20 or 21 wherein the braking distance of the second train includes a distance calculated based on the speed of the second train and a predetermined reaction time of a driver of the second train.
23. The method of any one of claims 20-22 wherein the predetermined threshold is at least 10 meters. 5 24.
The method of any one of claims 20-23 wherein the braking distance of the first train is determined according to the equation DL = -v02/233 and braking distance of second train is determined according to the equation DF = viTi + (-v12/2a1), where vo is the speed of the first train, v1 is the speed of the second train, ao is the braking deceleration of the first train, al is the braking deceleration of the second train and T1 is the reaction time of a driver of the second train.
A server comprising a processor and a machine readable storage medium storing instructions executable by the processor to: receive data relating to positions and speeds of a plurality of trains from on-board systems of the plurality of trains; identify a first train of the plurality of trains as a collision risk candidate in respect of a second train of the plurality of trains; and notify an on-board system of the second train that the first train is a collision risk candidate in respect of the second train.
26 The server of claim 25 wherein the instructions to notify the on-board system of the second train include instructions to send a message including a position and speed of the first train to the on-board system of the second train.
27. The server of claim 25 or claim 26 wherein the instructions include instructions to identify a collision risk candidate according to the method of 8, 9 or 14.
28. The server of claim 25 or 26 wherein the server is configured to perform the method of any one of claims 1-24.
29 An on-board system of a train comprising: a GPS receiver to generate data relating to a position of the train; an odometer to generate data relating to a speed of the train; and a controller configured to: transmit data relating to the position and speed of the train; receive data relating to the position and speed of another train; determine whether predetermined collision risk criteria are met based on the position and/or speed of the train and the position and/or speed of the another train; and cause the train to activate an event in response to determining that the predetermined collision risk criteria are met.
30. The on-board system of claim 28 wherein the controller is configured to transmit the data relating to the position and speed of the train to a remote server and receive the data relating to the position and speed of another train from a remote server.
31. The on-board system of claim 28 wherein the controller is configured to transmit the data relating to the position and speed of the train to the on-board system of the another train and to receive the data relating to the position and speed of the another train from the on-board system of the another train.
32. The on-board system of any one of claims 28-30 wherein the controller is configured to perform the method of any one of claims 1-24.
33. A system for monitoring trains comprising one or more apparatus which are configured to perform the method of any one of claims 1-24.
34. The system of claim 33 wherein the system comprises a server according to any one of claims 25-28 and a plurality of trains, each train comprising an on-board system according to claim 29 or 30.
35. The system of claim 33 wherein the system includes a plurality of trains, each train comprising an on-board system according to claim 31.
36. The system of any one of claims 33-35 wherein the plurality of trains includes a first train and a second train, the on-board system of the first train comprises a first GPS receiver and a first odometer to generate data relating to the position and speed of the first train and the on-board system of the second train comprises a second GPS receiver and a second odometer to generate data relating to the position and speed of the second train.