In news that will bring good news to many commuters, the first algorithm to detect delays automatically on the tracks has been developed. The intelligent tool was developed by researchers at the University of Portsmouth in conjunction with First MTR South Western Railway (SWR) via a two-year knowledge transfer partnership (KTP), funded by Innovate UK.
With more than 1,700 trains operating on SWR’s rail network across Southern England daily, minimising disruption to rail travel is challenging. It is difficult for controllers to detect delays promptly, which leads to further delays in selecting contingency plans.
SWR is a joint venture between FirstGroup and MTR Europe, two of the world’s leading train companies. With approximately 235 million passenger journeys a year, the SWR franchise is the largest and busiest in the UK railway network.
Despite large increases in passenger numbers, trains and crews, rail operators have been using the same systems and technology for decades.
“As a commuter myself, I’m delighted to be able to contribute to this project that will improve the customer experience,” said Dr Edward Smart, KTP Academic Supervisor. “It highlights the impact that machine learning algorithms can have for real world applications.”
University researchers automatically analysed data to determine the point of delay, identify which trains would be affected and select the appropriate contingency plans to get the services back on track. The intelligent tool is designed with machine learning techniques to significantly reduce the time to analyse and process the data.
Professor Chris Simms, KTP Academic Lead, said: “Automatic detection of delays represents the future of the rail sector. This project has made an important first step in realising the potential represented by machine learning to mitigate railway delays.”
The tool is currently being used within the SWR Control Centre, which is responsible for controlling the movement of trains across the network.
"Working with the University of Portsmouth has been an excellent experience for SWR and has transferred understanding into the business on systems development and AI,” explained Chris Prior, Head of Control Projects at SWR. “Together we have developed a system which improves the speed to response to recover late running, learn from and continuously improve SWR customer’s experience.”