Artificial intelligence being used to make transport safer and better

Stand at a particular pedestrian crossing along Paisley Road West in Glasgow, and you are being anonymously counted. One set of sensors monitors the number of people waiting to cross the road, while another set monitors approaching buses. The information is then being used to balance the priority of road users, taking into account the number of people waiting to cross and the proximity of buses.

This innovative use of technology is part of a partnership between Glasgow City Council and technology firm VivaCity. Known as Smart Traffic Sensors, the programme combines computer vision cameras with artificial intelligence (AI) to classify different types of road users. This can then be used to provide information and insight to local transport teams to help inform decisions.

“Our goal has always been to empower local and transport authorities with the data they need to optimise their transport networks,” says Graeme Cade, VivaCity’s managing director. “We are incredibly proud to collaborate with councils to help them achieve their aims, and Glasgow City Council is a great example of how forward-thinking approaches can be applied to transport planning. Their willingness to champion modern technology is exactly what is needed to drive progress and ensure transport in our cities becomes increasingly sustainable and efficient.”

Elsewhere in the city, VivaCity sensors are helping to monitor traffic, measure the success of active travel and road safety schemes, and optimising traffic signals in real time rather than relying on historic data. It has also been involved in tracking the effectiveness of a new AI traffic signal control system on Pollokshaws Road, a small-scale trial conducted by the council and tech firm SimplifAI in 2024. Analysis found the traffic signals improved journey times for buses by up to 50 per cent.

Pleased with the results, the council has now announced a larger one-year pilot of AI-powered traffic signals along the length of the road. Backed by £490,000 from the Scottish Government’s Bus Infrastructure Fund, the technology will, it is hoped, ease congestion along the corridor by modelling bus movements and enabling early intervention at peak times. The priority is on speeding up bus journeys, but the council expects all road users to benefit from better traffic flow.

It is one of the Scottish Government’s biggest investments in AI in transport to date. Speaking at the time the funding was announced, then connectivity minister Jim Fairlie said: “I recently visited Glasgow’s Operations Centre at Eastgate to learn more about the transformational impact that artificial intelligence and machine learning will bring to bus prioritisation across the city.

“Our investment will help improve bus infrastructure and speed up bus journeys, making buses even more attractive for people working, living in or visiting Glasgow. For our communities and for our climate, this crucial investment will support a shift away from cars and towards more sustainable bus services, by making public transport a more attractive and accessible transport option for even more people.”

Work to install the new traffic signals is still ongoing, but all these projects serve to highlight some of the ways AI and other new technologies are being deployed in Scotland to improve the public transport offering.

There is also significant research being undertaken at Scottish universities, partnering with stakeholders from across the UK. The TransiT research hub, for example, is focused on how digital twins can be used to help decarbonise the transport sector, which remains one of the most polluting, with improvements to public transport an obvious target.

Digital twins are virtual replicas of physical infrastructure, using data collected from across the transport sector. But more than simply models, they can be used to test ideas by using real-time data to show the probable impact of any decision if it were to be taken in the real-world.

Dr Joe Preece, a research fellow at the University of Glasgow, is working on a digital twin of public transport in Birmingham for TransiT. The specific aim of the project is to incentivise people to use public transport to the city’s Queen Elizabeth Hospital. “What we want to try and capture is the real-time state of the buses, where they are in the world, the traffic, the roadworks that are happening, the people moving in these systems, the weather – all of these and how they may affect particular decisions that we want to make,” he says.

The digital twin can then be used to test different options for altering people’s behaviour, for a fraction of the price of physical trials. “We could maybe introduce new bus routes, increase the number of buses running, we could look at where we put charging infrastructure for electric vehicles, we can play about with rotas and aligning train timetables with bus timetables; the possibilities are broad. As long as we define the problem clearly, we can try and test this in our virtual environment before we deploy it to the real world,” Dr Preece explains.

Another TransiT project involves the creation of a personalised digital twin assistant for passengers to use and plan journeys. Such technology would take into account mobility needs, journey requirements and personal preferences – and map out real-time options for travel.

Elsewhere, a research project from Loughborough University’s Transport AI Innovation Centre (Traice) is testing how AI could be utilised to make public transport more appealing.

Dr Haitao He, director of Traice, explains how sensors have been used to measure the level of comfort for passengers. “What we did is we installed lots of diverse sensors in terms of sensing the vibration, the temperature, the accelerometer in terms of whether they are jerks, in understanding how comfortable a bus ride is and why. It has been just a pilot test deployed in Loughborough, but we can clearly see that by collecting the data and using AI to analyse it, we can very well quantify the level of service and comfort for the customer, which I think will play an important role in guaranteeing that the people are getting the service that they’re paying for and that’s deserved.”

Dr He says one of the most common barriers for people not making more use of public transport is because it is not comfortable. The other, he said, is that services are too infrequent. AI could also help here. He points to the pilot of self-driving buses in Edinburgh which proved such vehicles are technically ready for deployment, but there remains a gap on the commercial side because there is not currently a “viable business model that can keep these technologies in operation”.

“One of the key elements that AI then can help with developing business model is in terms of predicting where the transport demand is and how it is evolving, both in terms of day-to-day from Monday to Sunday, but also the more longer-term seasonal and yearly variation,” Dr He says. “Why is this so important? We’ve run another research project at Loughborough University that’s looking at demand-responsive transport, which is facing a very similar challenge as those autonomous buses in terms of a viable financial and business model to keep in operation. This can be very expensive if you’re not running the routes at the right time and at the right frequency.

“We have shown in our research that by using AI that enables this kind of demand forecasting, you can deploy these demand-responsive buses – whether autonomous or not – much more efficiently. We can demonstrate that there is indeed a viable business model if you can get the demand right. That’s where I see the next technology that will truly enable these demand responsive processes.”

This speaks to a wider problem shared by many of these new technologies. For all the pilots taking place across the country, few are moving beyond these early stages. The UK, Dr He says, is “world-leading” in developing AI technologies and innovations in transport. “Where the UK is lacking or slightly behind is a stage to move from pilot to commercially viable deployment and operation and the scale up,” he says.

He believes there must be urgent action to rectify this issue, such as using procurement as a “strategic lever” to bring new technologies forward. “At the moment, most public sector procurement very much favours low-risk and existing solutions because the public sector is not the best at taking risks in technologies. In order to make some of those technologically viable in an internationally competitive landscape, I think we need to step up in terms of a more active procurement strategy that actively encourages innovation, like AI innovation. In general, we should be more innovative in procurement, particularly in terms of AI – that’s going to be very important.”

Ahead of the recent Scottish Parliament election, the Chartered Institution of Highways and Transportation called on political parties to make better use of data and AI for the sector. The organisation’s manifesto said: “Political leadership is needed to provide good governance of the types, quality and availability of data capture, and machine learning and AI developments.”

The UK Government published an action plan last year setting out how the Department for Transport (DfT) will work with the sector on AI. Its four broad objectives are adopting AI responsibly, maximising the economic benefits of AI applications, putting the UK at the forefront of transport-related AI, and using AI to drive quality, efficiency and effectiveness within DfT.

But transport is a devolved policy area, and Transport Scotland – the DfT’s equivalent – has no similar strategy in place. The Scottish Government’s broader AI strategy mentions transport just twice, only to acknowledge the need to create an environment where it can be adopted confidently by the sector.

The Scottish Government’s experiments with AI in transport therefore remain at an early stage, with the Glasgow bus trials the foremost example. While that does offer some insight in the direction of travel, far more will be required to deliver the whole range of solutions AI has to offer.

But as Dr Preece says, the future of transport is not just in AI technology. “AI is an enabler for other things, but it is one of a few tools that sit as part of a bigger system. There are other technologies that we’re using that aren’t under the banner of AI, but they’re still exciting and cool.”