No matter how good the planning – train delays often cannot be avoided for a variety of reasons, including weather conditions, construction works or malfunctions on the train path or vehicle. Sometimes a train may start later than planned because a track is momentarily blocked, for example.
A one or two-minute delay to the departure may not sound like much. But a tight train schedule often doesn’t allow lost time to be made up. So even just a few minutes delay at the start can block a track for a later train for longer and lead to knock-on effects.
At the ten key hubs of the railway network, the PlanStart programme from Deutsche Bahn AG is hot on the heels of such delayed departures. Employees work with the help of an app for tablets and smartphones.
Measuring points record every section when the train is prepared for service.
Approximately 400 train movements per day are currently being monitored with PlanStart. The best way to visualise this vehicle tracking is to accompany a train from its preparation for service at the depot all the way to the station. Along its route, the train passes several measuring points, whose passing is transmitted to the central system. Here the scheduled and actual times are compared in real time.
The first measuring point is directly at the stabling location or in the depot and has to be activated manually. For this purpose, DB Systel has developed an Android app that is operated on a smartphone or tablet in a relatively easy way: the dispatcher presses a button when the train is ready to depart and can also enter a reason for any delay to the train’s departure. In doing so, the dispatcher sets a digital time stamp and indicates possible causes for delay, which are then transferred to a database for further analysis.
As soon as the train starts its shunting movement, the GPS signal is used for precise localisation. This signal is updated every 10 to 30 seconds, and when the train arrives at the platform, GPS is used to detect when the relevant measuring point has been passed. A few minutes prior to the departure, the train manager issues a completion-message manually, again using the app. This provides the PlanStart team with the necessary basic data for analysing the causes of delays.
Interaction of GPS and manual input
Although the pure time stamp data recorded using the app is not very extensive, the mass of small data records that accumulates for around 400 trains every day, combined with existing data from other recording mechanisms, allows for “big data” evaluations. Hence the diversity of information is what’s important, not the actual size of each data record.
Simple app, big impact
Before the app was developed, one question was asked: how can we record data more efficiently? As every dispatcher and train driver carries a tablet and every train manager a smartphone, an app was the obvious solution. Within four weeks, the prototype was developed at DB Systel in coordination with the users, and two weeks later, the app was ready for use.
In actual fact, the app is just a virtual button that transmits a signal via the mobile telephony network to a server and also records reasons for a possible delay. Especially when combined with other recorded data, this information produces a punctuality forecast that is even more precise. For the PlanStart project, this is a practical addition for the aim of optimising the entire process, from preparation for service to the departure signal, at the ten busiest departure stations in the long-distance network.
Despite the short development time, the PlanStart app is functioning well. Since the start of the project, noticeable improvements have been achieved for passengers, and the results are helping to optimise the PlanStart system further. Yet the app is only an intermediate step until an automated system independently combines the data streams efficiently. Work on such a data analysis is continuing behind the scenes. And who knows, perhaps the experiences gained with the app will one day enable all measuring points to be recorded automatically. But that’s still in the distant future. Until then, a convenient app is impressively demonstrating how well highly developed technology with a human factor can work.