Three buttons sit on the driver's panel of Thameslink's Class 700 trains. After a brief stop in a new tunnel between Kings Cross and St Pancras, the driver of the 09:46 from Peterborough to Horsham presses the middle yellow button and the train starts to drive itself.
The journey was the first on the UK's mainline rail network to use an automated driving mode and sets the scene for thousands of partly automated journeys in 2019. "This is the biggest timetable change there has been in a generation," says Gerry McFadden, engineering director at Govia Thameslink Railway (GTR), the firm behind the project.
From May next year trains running through central London's Thameslink route will use the Automatic Train Operation (ATO) system. When turned on, it removes control of acceleration and braking from the driver – although this can be overridden – allowing the trains to run with smaller gaps between them.
One major benefit: automation will allow trains to carry more commuters – even if they have to stand. As a result it will mean that instead of 14 trains per hour travelling through central London there will be 24 per hour by December 2019, GTR says. Based on the company's figures it estimates the extra trains will create space for 60,000 more passengers by the end of next year.
McFadden says the ATO system will become standard across the UK within the next two decades. In September 2017 the government announced an extra £5 million of funding for National Rail to improve upon digital signalling.
The automatic driving works by giving control of the train's speed to a digital system. In many ways it is similar to the how the London Underground works: trains stop at the right position (they can be within +/- 50cm) on platforms and all the doors open at the same time.
Once the ATO is turned on it uses data from the European Train Control System (ETCS) to set the speed of the train. Using balises – a type of electronic beacon placed alongside rails – the ETCS transmits data to the trains passing above them. ETCS is essentially a digital signalling system, and was developed by Siemens for use in the UK, that makes the physical trackside signalling lights and indicators redundant.
But despite automation, McFadden says drivers are still needed. They are responsible for controlling the trains outside of London where the automated system does not yet work and for checking platforms are clear when the system is in use.
But the ATO can remove human inefficiencies. Jim Doughty, a system integration manager, says the biggest difference between automated train travel and that of humans is the speed the train approaches the platform. Doughty explains human drivers tend to stop accelerating when heading down a platform; they start "coasting," he says. During this time the train isn't travelling as quickly as it could be.
Doughty says the ATO only starts braking when it really needs to and does so smoothly, thus saving time. "No driver would drive like that," he says. As a result, trains can be run closer together. Both McFadden and Doughty say at their closest the trains using the ATO could operate around 100 metres apart.
"The actual system is designed for 30 trains an hour so it could always go above 24," Doughty says. "You're down to 2 minute 30 second headways [difference between trains] at 24 then two minute headways at 30 trains per hour."
The ATO system working alongside ETCS has been developed for more than two years. GTR, Network Rail and other partners have been testing trains running ATO in Germany, overnight on UK lines and at specialised test tracks around the country.
Onboard the 09:46 to Horsham, things go to plan. Once ATO has been activated the train pulls out of the Kings Cross tunnel, stops at the right place on the platform, and then once the driver has checked no more passengers are getting on pulls away again. This is a process repeated at Blackfriars station. "It's a performance tool," says the driver who has been pressing the start and stop buttons. And for regular passengers on-board, the idea is nobody notices whether human or machine is in control.
This article was originally published by WIRED UK
