With a growing number of new and disruptive modes of travel changing the way we move around UK cities, Katie Searles discovers whether traffic light signals can keep up…
Controlling how multiple modes of transport move through busy urban environments is no mean feat. This is only going to increase in difficulty as UK cities continue to become smarter. “It is hugely challenging in the face of a shifting city, it’s never the same,” says Irfan Shaffi, operational control manager, Transport for London (TfL).
It is perhaps no surprise, then, that the UK government has pledged some £15m to supplement existing local authority spending on the upgrade and maintenance of traffic signals and associated equipment.
This investment – to be spent in the 2021/22 financial year – is designed to enable the deployment of technology such as artificial intelligence (AI), machine learning, wide-angle cameras, and a range of sensors to help regulate traffic signals.
It is hoped that this latest cash injection will see smart traffic light systems installed across the UK, which are optimised to cut journey times, improve air quality and, as Shaffi describes it, “contribute to this humongous challenge – really to try and keep things going, keep things moving.”
When looking at which systems to introduce, there are a few questions that local authorities are asking. “Can we use cameras and machine learning capabilities to really understand what’s happening at a junction?” ponders Daniel Clarke, strategy and partnership, Smart Cambridge.
“Can we look down and see how many pedestrians there are? Are there cyclists? Is there a bus coming? How much traffic is there?”
“Then, because the camera can see all of the environment, can we then begin to manage it better and begin to prioritise different modes of transport – cycling or walking?” Clarke will be hoping to get answers to these questions following a 12-month trial with British AI transport company Vivacity Labs.
Currently, the majority of traffic signals across the city of Cambridge use fixed algorithms to make signal operation decisions. Vivacity’s machine-learning technology enables the new signals to learn what works best to manage traffic flows more effectively.
Furthermore, the camera-based AI sensors are said to be more able than existing systems to anonymously identify different types of road users and adjust traffic signal timings accordingly.
“Understanding how this kind of AI behaves in a wide variety of different environments and junctions is really crucial to being able to roll out on a large scale,” believes Mark Nicholson, CEO of Vivacity Labs. “One part of this is that we simply need to deploy it in more places to actually demonstrate to ourselves and others that this works really well.”
Vivacity Labs is currently working with around 60 different local authorities across the country, initially providing baseline data before exploring how that data is used in real-time systems, for real-time optimisation.
Another city that has deployed Vivacity’s technology is Manchester. “Being able to make use of AI presents a wealth of potential opportunities to the field of traffic signal control,” insists David Watts, intelligent systems engineer at Transport for Greater Manchester. “It makes use of deep neural networks and reinforcement learning to finely tune an extremely flexible algorithm – while retaining all safety critical features.
“By combining this with the vastly improved capability of video analytic sensors to detect active travel modes – and include these datasets as a primary input into such systems – this will result in a more dynamic and responsive algorithm, supporting increasingly important active travel networks,” Watts continues.
Using such algorithms to make data-driven decision on which traffic signals should be placed on green – and when – has the potential to determine the best way of moving not only traffic but also people around cities.
“Data is hugely valuable in managing traffic signals, particularly across a frequently changing and developing city region, where even relatively small changes can have an impact on congestion levels and travel behaviour,” Watts clarifies.
To test this theory, TfL recently teamed up with Siemens Mobility to use a real-time optimiser that takes into account a variety of rich data sources, covering all road users rather than just motor vehicles. It processes information on cycling, walking and freight movements across London.
However, as Shaffi explains, its the calibre of the data that is key. “We’re building an engine here,” he notes. “And the engine will only work if you’re giving it good quality fuel – good quality data.”
If the data-led system leads to a reduction in congestion, improvement in local air quality, and enables more people to get around the capital on foot and by bike, there could be a full-scale roll-out of the technology by 2023.
Getting the balance right
Expanding the use of such smart traffic signals has never been more timely since, over the past 18 months, the way people move around the UK’s cities has changed. An active travel agenda has been adopted by many local authorities and transport bodies. And, according to a recent UK Department for Transport report, this active travel push is working. Its figures reveal a 39% rise in people walking more and a 38% rise in those choosing to cycle. The same report suggests that 94% will continue to walk and cycle post lockdown.
But with more walkers and cyclists on the roads than ever before, safety is key. In many cases sharing the road space – especially at junctions – can be a fine balancing act. Traffic signal manufacturers are therefore looking to deploy smart solutions, AI and machine learning, to not only identify differing modes of transport but also to protect those exposed to the traffic.
In Newcastle city centre, two types of above-ground detectors have been installed by AGD Systems on cycle lanes approaching a junction to enable dedicated cycle phases.
“The increasing application of AI and machine learning within detection solutions to control traffic signals will further help to regulate crossing times for pedestrians and cyclists,” says Ian Hind, commercial director, AGD Systems.
“The technology has the capability to detect vulnerable road users approaching a crossing from further away and identify whether they’re going to cross the road or continue walking along the pavement.”
By protecting pedestrians and cyclists, these solutions will play their part in encouraging a switch to more sustainable modes of transport. This, in turn, should help cities tackle the air-quality challenge.
“The tools that we’re building have the capabilities to be entirely policy responsive and every city in the world is now waking up to the fact that the policies need to be shaped around dealing with environmental challenges,” Shaffi adds.
“The tools we’re providing now will categorically enable us to contribute towards improving air quality by making it easier, quicker and more convenient for people to use sustainable modes,” concludes Shaffi. “I’m not necessarily talking about penalising vehicle drivers but maybe starting to redress the balance because the vast majority of cyclists and pedestrians travelling around the city do see an imbalance there. We are very heavily motor dominated.”
“It’s working out what the trade-offs are between all those different modes,” agrees Clarke. “But by having better and more granular data on aspects such as movement and air quality, it allows us to have a much better understanding of the impact that those trade-offs have. And that can consequently help us to create much better policies.”
This balance can also ensure that – at the right times of day – vehicle traffic is the mode that is given priority. “Let’s say 8am, on a school day, we want to prioritise lorries because we don’t want them stopping outside schools,” Nicholson, explains. “We want to try to get them through the network quickly. Whereas at 5pm, you want to encourage everyone to cycle for commuting – so let’s try to get the cyclists through the network quickly at that time.
“It’s about giving them the policy choices of where they want to prioritise the different times, train the AI to react to those policy choices, and then deploy that in the real world.”
With a range of smart traffic light solutions being trialled across the country, it won’t be long before these intelligent systems of the future are commonplace. And, as Watt predicts, these lights will have “the capability to manage all parts of the network better, improving safety, efficiency and reducing carbon emissions.”
Work in progress
It’s not just permanent traffic light signals that are getting smarter – temporary systems are also evolving to keep up with developments in cities and to protect vulnerable road users.
For over a decade not much has changed in the way vehicles are controlled at traffic management sites. Many temporary traffic lights relay on Doppler radar for motion detection. But now companies such as Traffic Group Signals are creating solutions that harness frequency-modulated continuous wave radar to track more than just motion.
This solution can differentiate between large and small vehicles, it calculates the distance and speed of approaching vehicles and is also less susceptible to driving rain and non-target objects such as crisp packets in the wind.
“You get a much greater level of data accuracy and that makes it possible to control traffic in a more efficient way,” says Will Credicott, Traffic Group Signals.
Making these temporary solutions as efficient as possible is also important to SRL Traffic Systems. Managing director John Cleary explains that such systems can even be better than permanent lights. “That equipment will have been there for many years, and will not be as efficient as when it was first implemented, say, 20 years ago. When we come along with a full system, a bespoke layout, we validated it to that moment in time.
“We update timings, we make sure it’s working absolutely at peak efficiency for that junction, with many people feeding back to us that it’s working better than the permanent signals,” Cleary adds.
As well as ensuring temporary lights run as efficiently as possible, aiding with cutting congestion and therefore emissions, safety is fundamental for these suppliers. SRL’s temporary solutions harness technology to offer protection for cyclist and pedestrians.
It’s Urban64 intelligent temporary system has been installed at London’s Old Street works following concerns from Islington and Hackney Council that the roundabout was a collision hot spot.
To protect people using the four pedestrian subways and the area’s cyclists – which make up a third of all vehicles using the roundabout – SRL installed a two-way, signal-controlled layout following the closure of the north-western arm of the roundabout, creating a new peninsula space with the existing central-island.
“It was exciting to be able to fully use Urban64’s features of incorporating cycle phases and signals into a temporary system, and to see it working so successfully,” Cleary enthuses.
Additionally, Traffic Group Signals is also working to develop technology that enables the effective management of pedestrians in roadwork sites, eliminating the need for an operator on-site having to manually control the lights.
“We are striving to make our products safer, greener and more efficient,” adds Credicott. This includes developing a solution that improves driver behaviour. “Because the lights are performing better, a driver is less likely to become frustrated,” assesses Tom Miles, senior commercial manager at Traffic Group Signals. “So workers on site are less likely to have a car pulling down a closed lane.”
This not only benefits the driver but also increases the safety of the teams of road workers, which is paramount. “Technology like this can only benefit everyone in the long-run,” notes Miles.
It’s because of these facts that Miles feels the industry is waking up to the potential of smart solutions. “The industry is really starting to gather some speed in terms of what’s being required of them – not just ensuring efficient movement of traffic, but environmental factors and social responsibility.
“I think it’s going to drive a lot of change over the next decade or two. It will be a completely different place in 20 years from what it is now,” he concludes.