VIDEO: TransPennine Express tests UK-first AI speech-to-text announcements designed to improve journeys for hearing impaired passengers
A cutting-edge announcement system designed to improve journeys for hearing-impaired passengers on TransPennine Express (TPE) trains has successfully completed its pilot phase and is the first-of-its kind to be used on the railways in the UK.
Developed in partnership with tech company Televic, the new system allows passengers to read real-time transcriptions of live announcements made by conductors on TPE’s onboard screens. This is especially helpful for customers with hearing loss or even those wearing headphones - making sure they don’t miss important information.
Funded by Innovate UK, getting the system up and running wasn’t easy as the team had to overcome major technical challenges. The AI needed to understand a wide range of accents, work seamlessly while the train is in motion, and handle background noise on board—all without relying on an internet connection.
Kathryn O’Brien, Customer Experience & Transformation Director at TransPennine Express, said: “The speech-to-text announcement system complements the live train information already displayed on screens onboard our trains, such as arrival and departure times, by adding a personalised touch to our service disruption updates and other announcements.
“Our two-day trial was a huge step forward, showing that this AI-powered speech-to-text system has the potential to change the future of rail travel.
“We want our trains and stations to be accessible to everyone, and we hope this innovation will give people the confidence to travel with us – no matter their needs.”
With development still in the early stages, TPE will now work towards installing the innovation across its Class 397 fleet.
Steven Lauwereins, Research Lead at Televic, said: “Visualizing manual announcements is the next step in Televic’s goal of improving passenger experience and will be a game changer for accessibility. Our live transcription of announcement system uses innovative onboard transcription models to guarantee high quality transcriptions in real-time.”