Artificial Intelligence – or machine learning – is hard to avoid these days. It feels like every new piece of technology uses AI, including those that help improve road safety.

Australian company Acusensus is currently working with National Highways and consultancy Aecom to monitor driver behaviour across England’s motorways and major A-roads in Warwickshire, with its Heads-Up Realtime software.

A sensor-equipped van can be located anywhere on the road network, its multiple cameras able to individually capture every passing vehicle, even at speeds of up to 186 mph. The AI software can instantly determine whether drivers are using a handheld device or wearing a seatbelt.

On announcing the trial, National Highways head of road safety Jeremy Phillips said: “Sadly, there are still drivers who do not feel the need to wear a seatbelt, become distracted by their phones or travel too close to the vehicle in front. We want to see if we can change driver behaviour and, therefore, improve road safety for everyone.

Dr Jamie Uff, technical director for strategic transportation consultancy at Aecom, added: “Despite the often-reported dangers of distracted driving and failing to wear seat belts, the numbers of people killed or seriously injured as a result of these behaviours remain high. The technology we’re deploying makes detection straightforward and is providing valuable insight to the police and policymakers on the current level of road user behaviour. We are really keen to use this equipment to raise awareness and help improve road safety for all.”

As an indication of how important this pilot is, 23% of car occupant fatalities in the UK in 2020 were linked to not wearing a seatbelt. In that same year, 17 people were killed and 499 were seriously injured in collisions where the driver was using a mobile phone.

This isn’t the first use of AI as a road safety tool. Aecom previously worked with Acusensus, trialling a fixed camera in 2021 on the M4 in Berkshire that could detect drivers not wearing seatbelts and holding a mobile phone. In six months, the camera and software detected nearly 7,000 people failing to belt up and over 25,000 drivers holding their mobile phone.

National Highways has also been trialling new tailgating cameras on a stretch of the M1 near Northampton, capturing 60,343 instances of vehicles driving too close between October 2020 to September 2021.

AI has a bright future, as Nick Reed, chief road safety advisor at National Highways tells us: “While we wait for self-driving vehicles, there are many other uses of AI that could contribute to road safety. One is the ability to monitor, classify and track the movement of road users using machine learning. Once suitably trained and calibrated, systems today can use videos of traffic systems to track individual vehicles, cyclists and pedestrians automatically and even start to understand whether observed behaviours are of concern. This ability to use video streams to understand where risks are emerging on our roads could be a critical source of information that enables roads authorities to mitigate issues before collisions occur.”

The Acusensus trial is another milestone in the journey to make AI an integral part of road safety enforcement. Expect to see many more such trials, followed by rollout of systems over the next decade, as AI takes over the world.

Have you worked on a road safety project recently? Enter the CIHT Road Safety Award 2023 and get the recognition you deserve. Deadline for entries is Monday 6 March 2023.

(photo credit: Acusensus) 

Words by Craig Thomas

Have you worked on a road safety project recently? Enter the CIHT Road Safety Award 2023 and get the recognition you deserve. Deadline for entries is Monday 6 March 2023.