Dr. Claus Bahlmann tells us how Siemens Mobility is using artificial intelligence (AI) to automate railways, and why it will make rail more flexible, more efficient, and more inclusive.
With automation – powered by AI – we can increase flexibility to deal with varying peak travel times, reduce the downtime caused by maintenance and staffing issues, and even carry out upgrades without affecting rail operation so much. All of this benefits the passengers.Dr. Claus Bahlmann, Head of R&D Department Artificial Intelligence, and Principal Expert for AI and Computer Vision in Siemens Mobility
Population growth and urbanization are increasing pressure on public transport. How can rail systems cope with that increase in passengers?
It’s all about getting smarter and automating as much as possible. With automation – powered by AI – we can increase flexibility to deal with varying peak travel times, reduce the downtime caused by maintenance and staffing issues, and even carry out upgrades without affecting rail operation so much. All of this benefits the passengers.
There’s a limit to how much new railway infrastructure we can build, and investment is also limited, which is why we need to take this digital approach. The technology is there to increase capacity while making rail travel more reliable, more comfortable and more inclusive.
AI is key to automation, because of the complexity of transport systems and the world they operate in. If we want highly automated vehicles to navigate in that complex world, we have to be able to perceive what’s going on in the environment – the infrastructure, any obstacles along the track, the passenger load – and how the vehicle is responding.
How much of the rail system can be automated?
We aim to automate all areas of the rail system. However, the difficulty depends on how much interaction with the outside world there is in a certain system. Airport people movers or selected metros are already automated to the highest degree; here we talk about Grade-of-Automation Level 4. This is possible because they can be designed as closed systems. But with a regional commuter train there is always the possibility that somebody puts something on the tracks or kids are playing on the tracks. Then there are trams, which share the space with cars, pedestrians and cyclists. The more interference there is, the more challenging the automation.
So, to cope with these challenging scenarios, the rail industry has been starting to develop systems that deploy sensors on the vehicles and letting them perceive the environment and respond to it. We can look at the environment and at the vehicle itself and then apply intelligent algorithms to draw conclusions from what the vehicle sees, and then act on it. This development is currently in the R&D stage. It still brings a few technical challenges, in particular in the area of safety assurance. But it has enormous potential towards higher flexibility in rail transportation.
AI can also make maintenance more effective and efficient. Technicians in the field have traditionally relied on a service log and manual, but now they can point a tablet at the area of concern and AI will identify what part they are looking at. AI will then provide all the contextual information they need: the manual, the circuit diagrams, or a catalog of spare parts, which they can order from with one click – for instance in “Siemens Easy Spares”.
Inspection is another area where AI is being deployed into our products. When sensors are mounted on the infrastructure it is possible to inspect the rolling stock, because vehicles are scheduled to pass that location, and sensor data of the rolling stock can be collected and inspected. And when sensors are mounted on the vehicles, they will acquire sensor data from the infrastructure and the AI will recognize specific areas that need inspection or maintenance.
For example, we can inspect the condition of rail tracks from images captured by cameras mounted on trains and looking into the track bed. This technology has been rolled out in the “Railigent Video Track Inspector”.
Presumably that kind of automation improves the service for passengers. Does AI help with any other customer needs?
AI can also make rail travel more available to people who currently find it too difficult or unsafe. For example, someone using a wheelchair who wants to board a train often finds that the entrance is already blocked – either by another person with a wheelchair or just by too many people standing there.
We have developed technology that uses CCTV cameras mounted on the train to detect which areas are blocked and which are available, and this information can be sent to people waiting on the platform so they know where to stand. It can also aid social distancing in the pandemic: knowing which areas of a train there is high passenger density helps people on the platform to decide where to board the train. Of course, there is also a security benefit – the system can detect aggressive or violent behavior when it happens.
And once we get driverless vehicles on the tracks, it will open many more opportunities. That will be a game-changer. Once the train is driverless, we can move from a fixed transportation system with fixed schedules and fixed-capacity trains to something much more flexible – for instance, having a demand-driven train service in the suburbs or at night. Today, with trains operated by a driver, this is not possible due to the availability and schedules of train drivers, and for economic reasons.
And, going back to the inclusivity point, many elderly people don’t use public transport because they might need to wait for long periods – especially if they live in the suburbs. With driverless trains bringing more flexibility to the system, trains can operate more on demand, which will help those passengers.
Are there any projects where AI-led automation is already happening?
Yes, in addition to the “Easy Spares” and “Video Track Inspector” examples, Siemens Mobility uses AI to efficiently digitalize railway infrastructure together with our customers. This includes projects with DB and Bane NOR. The Bane NOR project to digitalize Norway’s entire railway infrastructure is scheduled for completion by 2034. The new centralized digital signaling system will enhance safety, punctuality, and capacity.
Siemens Mobility is building a digital copy – a ‘digital twin’ – of the central interlocking system and all infrastructure elements involved in this by using a combination of sensors and AI. A sophisticated sensor system with multiple cameras, LIDARs, and a positioning system, is being moved through an existing rail network to collect sensor data, such as camera images and LIDAR point cloud. Sophisticated AI is then used to analyze the data, creating a map of tracks and existing infrastructure elements accurate up to a few centimeters. This means we can avoid the laborious and time-consuming work of putting people into the field to carry out tasks such as measuring distances; we can simulate something in the digital world before we roll it out in the real world. Using our precise map of what is already in the field, we can start building the next generation of the system.
Have you changed the way you engage with customers?
I think there is a greater focus on co-development together with customers. We work with them to identify the pain points – the real issues. Then we have quicker innovation cycles in response to that engagement, and we validate the hypothesis in a short time. This is a new way to engage with customers, and I believe also with society on the big changes that come from AI.
