A new artificial intelligence app has been trialled on Highways England’s A14 Cambridge to Huntington improvement scheme, enabling ‘high risk’ days to be identified.
The app allowed staff to record both hazards and good practice, with the information then used to identify potentially challenging days. At its peak, some 4,500 observations were submitted in one month as part of the £70,000 initiative.
Information logged during the trial ranged from health and safety observations to information about vehicle movements and commercial activity. Daily risk profiles were then generated based on the data collected, identifying which days were higher risk and why, reports New Civil Engineer.
The AI identified that one of the most prominent themes affecting risk days was staff fatigue. This allowed practical measures, such as ensuring breaks are taken, to be implemented to improve safety.
Some of the key findings of the project that was found to increase risk included: working past 6pm, working more than nine hours, working in high winds and working following heavy rainfalls.
The AI pilot achieved 75% success in identifying high risk days. An unexpected outcome was the discovery that a workforce engaged in reporting hazards results in less harm. Therefore as the project progressed and incidents of harm were reduced, the accuracy of the model reduced to 65% as there was less data for it to analyse. Despite this, it still proved 160% more accurate than merely guessing which days were likely to have an increased risk.
Highways England’s A14 Project Director David Bray said: “Highways England has adopted the A14’s approach and technologies for its own digital platform which is being rolled out across our portfolio,” he said.
“We believe this technology will deliver significant cost savings as paper-based reporting is replaced by real time data, consistently captured and accessible to our project managers.”
Highways England A14 Head of Project Management, Mark Tootell, who led the project, said: “This exciting research demonstrates three important principles that projects should adopt: the importance of fostering a healthy observation reporting culture across the entire workforce, the value in taking a variety of factors into account and the need for well organised, broad and honest data sets.”
A14 data analyst Jaydip Jani added: “The AI health and safety project showed incredible promise and will only improve as more, better quality data is available. It has shown the industry the intrinsic value in well-gathered data and the potential for truly data-guided decision making.”