DfT harnesses Gaist’s AI expertise for roads audit as it gears up for driverless vehicle revolution


UK, 20 July, 2019
– Computer-vision technology and AI is to be used by the Government to assess the quality of the country’s road markings amid warnings that the poor quality of the local network could stall the introduction of driverless vehicles.

A £2 million audit of national transport infrastructure by the Department for Transport (DfT) will include an assessment of white lines and other road markings as well as a health check of sections of the National Cycle Network and the country’s footways.

The audit, which comes as the country gears-up for the introduction of driverless vehicles, will use AI to enable vast volumes of road data to be analysed and assessed.

The Government has committed to having fully self-driving vehicles on UK roads by 2021 as part of its Industrial Strategy. However, studies from organisations including the RAC, have warned of the need for a comprehensive review of roads infrastructure and an audit of likely maintenance and renewal requirements before the full launch of Connected and Autonomous Vehicles (CAVs) can get underway.

The DfT will execute the work in partnership with the Local Condition Roads Innovation Group (LCRIG). The technology is being supplied by Gaist, a North Yorkshire based SME, which specialises in the surveying of critical infrastructure and its analysis using AI and other data-science techniques. The audit which will form the largest, most comprehensive assessment of road markings ever undertaken.

Gaist will use images drawn from its national databank – a ‘digital twin’ of the local road network which it has developed over six years and which contains more than 1.8 billion images covering 640,000km of roads. For this DfT project, Gaist will draw on more than 146 million of its images, covering 96,000 miles of classified roads.

The first AI survey of its type, will be undertaken at Gaist’s AI Hub. It will enable the government to gain a comprehensive picture of local road markings across Britain  and to assess where improvement is most needed to ensure that safety is not compromised. Self -driving cars understand where they need to drive and that passenger safety will not be put at risk.

Chris Grayling, the Transport Secretary, said: “Road markings play a vital role in keeping everyone who is using the road safe, so making sure they are up to standard is imperative.

“This funding will allow for advanced AI learning technology to assess the condition of the markings to improve the safety of our roads for all users.”

Paula Claytonsmith, MD of Gaist, said: “We are using over 146 million HD road images from our national databank and cutting-edge AI technology to assess over 96,000 miles of classified roads as part of this project. This is the largest exercise in assessing road marking readiness ever undertaken in England. Gaist is proud to have the AI capability that puts an SME UK business at the forefront of technological advances.”

The Government audit follows publication in 2017 of an RAC Foundation report which explored the readiness of the road network for autonomous vehicles.

It explained that “autonomy could require enhanced standards of road maintenance, to ensure that driverless vehicles are able to ‘sense’ the road environment accurately – the delineation of the carriageway, lane markings, traffic signs and signals.”

It added that “whatever trajectory emerges for the driverless car, we won’t be travelling very far unless we have an adequately maintained network to drive on.”

The poor condition of Britain’s local road network came under the spotlight recently, when the Transport Select Committee published its local roads funding report.

It cited a study from the Asphalt Industry Alliance this year that showed 11 per cent of all local roads were in a poor condition and a further 25 per cent showed deterioration. The report concluded that the repairs would need £10 million to address.