Track vacancy / train detection

Notes - to be moved to proper section in feature paper (e.g. Train Movement), ref feature_movement

Track vacancy

Track vacancy - i.e. whether a specific section is clear and can be used by a train or is occupied and cannot be used - can be detected in may ways. Traditional interlocking is using track circuits, axle counters and the like. If reliable position reports can be achieved from all trains in an area, this can also be used to determine if a section is occupied or clear. Actual train position is traditionally determined by odometry (measurement of travel distance) based on “mile stones” readable by the train.

In the WinterTrain project track vacancy will be determined solely based on all trains regularly reporting their position. At regular intervals all over the track network so called balises will be installed. A balise is a small passive transponder mounted in the track. When a train passes a balise the balise will transmit a unique ID to the train. Starting at the position of the balise the train will measure - as it moves along the track - the distance between the balise and the train.

At regular intervals each train will transmit a position report consisting of the actual distance between the train and the last read balise, the ID of that balise and an indication of the driving direction.

The RBC will have a data model representing the track network including location of all balises and relative distances between balises and other elements of the network. This model will as well include information of the length of each train and the position of the balise antenna at each train.

Based on this model, train data and received position reports from all trains, the RBC will be able to determine where in the network each train is located and hence which parts of the track network are occupied and which are clear.


Given a position report and the model of the track network as the only information the location of a train will in most cases be unambiguous. That is, the RBC can determine the location based on the report and the model alone.

One exception exists however. If the track network between the last read balise and the train includes a facing point then the location becomes ambiguous. A point is called facing when the train is passing the point in direction from tip to branch. The problem is that the train has no knowledge of which of the two branches of the point is it moving along. In this situation knowledge of the previous location of the train will be needed in order to unambiguously determine the location of the train.

If the previous location is unknown (e.g. due to reboot of the RBC) and current location of a train is ambiguous the RBC has to assume that the train is occupying both branches of the point.

it/wintertrainv5/rbc_notes/tvd.txt · Last modified: 2021/05/01 08:36 by jabe