Are smart junctions just around the corner?

Artificial intelligence (AI) powered ‘smart’ junctions might sound like something from the distant future, or an unproven concept fresh from the lab, but in reality that’s far from the truth. In fact, AI is already beginning to play a role on streets and intersections across the globe. 

Growing urban population density

While the relative maturity of the market might be surprising, the key global driver will be recognisable to all - the exponential growth in city populations. In Europe, the majority of people live in urban environments, with more than 60% living in substantial (over 10,000 inhabitants) urban areas. 

One of the downsides to this is pollution, with urban mobility accounting for 40% of all CO2 emissions of road transport and up to 70% of other pollutants from transport - a challenge that the growing electric power train movement is a long way from effectively tackling. That pollution can spike for a variety of reasons, but congestion is one of the biggest contributors - indeed, traffic congestion costs EU economies €100 billion ($118 billion) annually. 

Global trend lifts all ships

It’s not just a European issue, of course, with figures from across the world showing a similar trend - a recent survey by traffic analytics company INRIX found that the total cost of lost productivity due to congestion in the US stood at $88 billion in 2019. Within that big number, there are many smaller tales of woe - for example, in the most congested city in the US, Boston, the average commuter in the metro area loses 149 hours per year to congestion, costing $2,205 per driver. At the other end of the scale, Wichita, Kansas sees drivers lose less than two hours a year.

Given those drivers, the rate of adoption of smart city technologies looks much less surprising, although the list of cities is indeed varied, from Mumbai to Buenos Aires and Madrid, Manchester to Cambridge, smart junction technology is out there. 

Substantial impact on city life

One manufacturer, Kapsch TrafficCom, supplies ‘adaptive’ traffic light management systems that can respond to developing situations and vary traffic patterns accordingly, with some success - the company claims that a Madrid installation saw traffic jams reduced by around 20%. The system links roadside sensors, traffic cameras and vehicles that monitor the various congestion factors such as accidents or road works in real time, sending a flow of data back to the centralised AI, which in turn models the traffic flow. 

That data can then be used to subtly manipulate traffic light signals to modulate flow. The company also delivers smart congestion charge technologies, which reduce the impact of the traditional ‘rush hour’ by incentivising drivers to avoid peak times. 

Is 5G set to widen adoption?

Manchester City Council in the United Kingdom is set to invest in a £2.3mn project that aims to deploy smart junctions through 2021 via the Department for Digital, Culture, Media, and Sport’s ‘5G Create fund’. The project in partnership with Vivacity Labs, will introduce AI-controlled ‘smart’ traffic junctions to reflect and accommodate an increase in cycling and walking in the city. 

The system relies on sensors with inbuilt AI - a scenario assisted by 5G connectivity - which detect different transport types and prioritise them differently at junctions according to the wider traffic flow picture. Rollout is underway, with three neighbouring junctions in Salford under Vivacity Labs‘ AI control in September 2020. Transport for Greater Manchester (TfGM) and Vivacity believe the system will cut emissions and improve air quality, as well as reduce congestion and queuing. 

Manchester is far from alone in seeking solutions to these globally resonating issues - a separate project in Cambridge will see a trial of four junctions connected with a Vivacity camera-based AI system, also targeting different types of road user. The 12-month trial funded by the Greater Cambridge Partnership (GCP) is scheduled to begin in summer 2021 and initial findings will be available in Spring 2022.

Pandemic data blip challenge 

Ironically, a key inhibitor that has emerged to prevent far greater rollout of smart junction technology is the COVID pandemic - not for the usual reasons of global supply chain delay and localised lockdowns - but because congestion has been mitigated at a stroke.

Albeit a highly temporary situation, but the latest INRIX figures show a vast decrease in traffic volumes. For example, in the UK alone, the average person spent just 37 hours stuck in traffic in 2020, down from 115 hours in 2019 – a decrease of 68%. This certainly has the potential to disrupt pilot projects targeting a congestion metric over the coming months, as well as related pollution-reduction projects. 

The use of AI to smooth out real-world traffic issues is clearly set to increase exponentially over the coming years, especially when considered as part of an overall trend to limit pollution (especially particulate pollution) in cities across the globe. 

Existing standards such as V2X technology are likely to play a role too, bearing some of the integration load and creating a ready network of ‘smart’ devices, from lamp-posts to traffic lights, road signage and vehicles. The use of smarter ‘congestion charge’ techniques to incentivise drivers to avoid peak times and places is also likely to see increased uptake, as governments look to tackle congestion-related waste and encourage more sustainable transport methods such as cycling and walking. Indeed, from a social perspective the result of increasingly sophisticated smart traffic management may well prove to be a rise in the simplest transport methods around…