Though it is now somewhat less of a bold vision than when first brought to the public’s attention, artificial intelligence (AI) – in short, computers making decisions that would otherwise be the preserve of humans – has very much arrived and the smart money says it will very much be part of the future. However, it is one thing AI governing the quality of a photo as it does on Huawei’s latest phone, or playing chess guided by the power of Google’s DeepMind operating system, but quite another when life itself could literally depend on it. That’s certainly the case in transport and while 2013's and 2016's crashes in Santiago de Compostela, Spain, and Bavaria, Germany, are incredibly rare, errors can prove catastrophic. And while you will see here that AI-enabled systems do indeed have the power to control vehicles, we also shine a light on the other forms of tech that don't have such a heavy burden of responsibility on them, protecting revenue and reliability rather than against the threat of a major emergency.
Launched officially in Potsdam during last year’s nearby InnoTrans show, Siemens showed what could be achieved by AI with a new breed of tram that can avoid the threat faced by tram drivers on a daily basis: people walking in front of a moving vehicle. The German manufacturer’s AI Combino tram can come to an abrupt stop should something cross its path thanks to a host of sensors, lasers and cameras that help make the kind of split-second decisions that elude humans. It’s still in development phase but it surely has a great deal to offer the wider autonomous vehicle market that is building pace at an impressive rate of knots.
Tickets and booking
Trawling every one of the roughly 500 million daily tweets for useful information is a task beyond that of any human, which is why Trainline is using the power of AI to do it . The transport ticketing company is using a system to help make journeys smoother for passengers by aggregating rail disruption information from Twitter in order of importance. Trainline then takes that info and matches it with Google Assistant users asking for specific travel-related questions, such as: ‘Are there train delays tonight?’ or 'Is the weather affecting services?'
With other train-related tech that relies on the human voice, SmartRail World recently reported about Virgin Trains’ Amazon Alexa ticketing system service that enables those with special requirements to book not just tickets but also platform assistance. Though not singled out by the train operator for its AI capabilities, news from the shopping giant Amazon has separately launched Alexa Hunches, a feature that promises to use “deep neural networks and machine learning” to remind home owners to complete tasks that may have slipped their memory, such as locking the front door or turning off lights. So, perhaps that signals further room for improving the voice-activated service.
To learn more about the kind of technologies mentioned in this feature, SmartRail World is hosting a range of events in Europe and the US. The next, SafeRail, is taking place in Washington D.C. on May 14th-15th. To read the agenda, see the speakers taking part and to register for the event, visit the show website.
Meanwhile in India, developers there have also harnessed the machine learning capabilities of AI to bring about a customer service chatbot based on the technology that will reportedly be capable of conducting conversations and answering questions related to travel bookings. What’s more, the service known as Ask Disha, which is being billed as the answer to cutting the country’s infamously difficult-to-navigate queuing system at stations, will also be conversant in 24 different languages. According to press in the country, Ask Disha won’t be able to answer specific enquiries relating to delayed services like Trainline’s system, but will direct users to the correct URL which can. The joint enterprise between the Indian Railway Catering and Tourism Corporation (IRCTC) and the Bangalore tech company, CoRover Pvt, will be integrated with integrating Ask Disha with ICRTC’s bespoke Android-specific app.
Embracing the ever-important role that data can play in the smooth running of infrastructure, train leasing company Porterbrook has teamed up with the original masters of IT, Microsoft, to improve rolling stock operations. Also working with the data science consultancy, Elastacloud, Porterbrook will use the large volumes of data that Porterbrook admits can all too often be ignored. According to the two companies, around thirty key performance indicators have been identified across the rolling stock industry that might be improved through the use of AI, using it to intelligently analyse data to help drive real improvements to fleets and, in turn, passenger journeys.
That same approach of using AI-backed tech to keep infrastructure operate optimally in also employed in India, where developers have created a system that takes photos and videos of the underside of trains to inform maintenance teams of problems areas before they are exacerbated. Capturing high-definition footage of vital parts of the train, such as gearing, the cameras are able to get into nooks and crannies that would be simply impossible for a human workforce. Used by India’s Central Railway, once full testing has been completed, the Undergear Surveillance Through Artificial Intelligence Assisted Droid (USTAAD) could be rolled across its operations.
This may also interest you: