SIEMENS

As Peter and Inge leave their house in the morning, they have their day carefully planned. Peter has to go into town for a meeting, and Inge is going to spend a few days with a friend. Peter wants to travel the last few kilometers by tram, but his smartphone tells him that just a few minutes ago a truck collided with a streetcar and knocked it off the rails. As it turns out, the high-speed train that Inge has booked is also stuck – behind a downed overhead powerline. Luckily for them, the rail operator has provided alternative travel options along with information regarding the delays.

Peter’s route involves a detour via subway and multiple transfers. Inge has somewhat more time. She decides not to board the next train, which is already overcrowded, but rather the one after that. For her cooperation, she will be compensated financially by the rail operator and will get a free breakfast. Peter and Inge both arrive at their destinations, though somewhat later then planned. However, the consequences are fairly minor, because they were informed in a timely manner and the rail operator was able to minimize the effects of the disruption.

Right now there isn’t a train operator anywhere in the world that offers a delay and planning app like this one for its customers. The databases are not suitably networked. Instead, for historical reasons, they have developed independently. But the ability to react quickly and flexibly would certainly be worthwhile. It would save time and stress for travelers, and money for rail system operators.

Siemens engineers and technicians have been working for years to take fault and emergency management, which is still fairly jerry-rigged, into the IT era and place it on a firm mathematical foundation. At the moment, dealing with broken rails, snowstorms and defective locomotives comes down to experience and having a nose for what to do.

Each disruption of a rail network reduces its carrying capacity. Storm damage, for example, can suddenly transform a high-speed two-track segment into a single-track segment. The dispatchers then have to manage the remaining capacity carefully. Intuitively, probably everyone would assign two full commuter trains a higher priority than a long-haul train that is less full. However, that would be a mistake if a lot of those passengers had to be put up in expensive hotel rooms. Depending on the situation, the nonintuitive option might be the correct one.

In any case, there will no longer be any trains that automatically have the right of way, as was the case in the 1970s. Back in those days, long-distance traffic always had priority over local service. Today, even highspeed trains wait if this helps the operator to avoid incurring significant financial penalties for delaying local commuter traffic. The dispatchers therefore have to analyze the technical and business aspects of each case.

Instant Optimization. “What we have here is a classic optimization problem,” says Dr. Stefan Wegele, a mathematician at Siemens Rail Automation. “We are looking for the best possible answer to a given problem.” To achieve that goal, automatic processes must analyze the prevailing parameters and generate all the solution scenarios available to the railroad for quickly getting as many travelers as possible moving again. A program works its way forward step by step to the right solution. It uses an abstract model of the situation in the rail network, which includes all available track-and-train configurations, as well as the number of passengers to be transported.

In addition to empirical considerations, the program takes economic parameters into account. These are specific to each customer and describe the customer’s business model and contractual situation. They include contractual penalties to be paid to the state or local government in the event of delays. Siemens specialists incorporate these factors into their algorithms. Software takes all of this information into account to find the best possible strategy for managing a given situation. After fewer than 100 iterations, the algorithms deliver the optimal solution in just a few seconds. “We are playing a chess game against chance, and just as in real life, the computer ultimately wins. That’s because it never gets tired, and it arrives at the right strategy for every constellation in a flash,” explains Wegele. “Our experience with railroad automation provides the starting values for all of this.” This experience goes way back. After all, Siemens has been manufacturing railroad automation and signaling systems in Braunschweig, Germany since 1873, and has been conducting research in this field for the same length of time.

The prerequisite for any decision is a complete and current model that perfectly describes the network situation. “But a look at the IT landscape used by rail operatorsaround the world shows isolated systems almost everywhere,” says Gerd Tasler, product manager for rail IT solutions at Siemens Rail Automation. “Operators often only have an overview of what is going on within their own area of responsibility. What’s missing is comprehensive networking and complete data integration.”

To overcome this situation, operators need a planning tool that includes all the economic parameters and technical data associated with their trains, such as top speed and performance. Train occupancy data are also important. More and more rail operators are therefore retrofitting sensors that provide information regarding the current load factor.

Rapid analysis and precise results are no guarantee for fast, appropriate decision-making. To make the results derived by a computer available to all members of a decisionmaking team, Siemens has developed the “multi-touch table,” a device something like an oversized tablet computer with touch screen technology. Introduced at the Innotrans trade show in Berlin in 2012, the device provides multiple persons with a shared, interactive interface for access to all information that is relevant for real-time rail operations management. A one-of-kind system, the table supports the emergency response team in its work.

No Instruction Required. During development of the device, a lot of emphasis was placed on intuitive operation. “Anyone who uses a smartphone will be instantly familiar with the interaction options we have provided,” says design specialist Kim Rosenthal of Rail Automation in Braunschweig, Germany. Test users mastered the use of the table within minutes. For example, spreading your fingers apart or squeezing them together either enlarges the image or reduces it to provide a better overview, similar to the one provided by a smartphone.

That eliminates the need for lengthy instruction and minimizes input errors. Four to five people can work on the table at the same time without overtaxing the system. During input, a series of infrared sensors record down to the millimeter where a finger touches the screen. When that happens, alternative routes and action options appear within seconds. “The latest-generation multitouch tables are very reliable and are ideally suited for use as large-area input and output devices for a team of decision-makers,” says Rosenthal. “Users tend to accept it as a tool very quickly. The gestures used become second nature in no time at all.”

The new technology will be rolled out gradually. “Components of the railroad IT system are already being implemented, such as in a major project in Copenhagen, where we are building a municipal rail line, and in New York City, where an existing subway is being upgraded with a modern display system called PACIS,” says Maximilian Eichhorn, Vice President, Rail IT Business, at Siemens Rail Automation. If everything stays on track, the delay app that got Inge and Peter to their destination with no stress might be available for download within the next few years.