Road Maintenance Robots

Sealing cracks before they become potholes

Some of the most practical road robotics work targets defects before they turn into holes. A Local Government Association case study describes Hertfordshire County Council working with a technology company and University of Liverpool researchers on a machine that uses scanning and AI to locate and assess cracks and other surface defects, then fills the cracks automatically to stop water getting in. The case study calls the council's trial the world's first public road trial of the technology. For a highway authority, the interesting question is less whether such a machine works once and more how its findings compare with your inspectors' records and how long its repairs last.

Taking crews off the cone-laying lorry

Placing and collecting cones by hand from the back of a moving lorry is among the riskier jobs on a road network. When Highways England trialled automated cone-laying vehicles in 2020, it said road workers could lift as much as 10 tonnes of equipment per shift, mostly at night beside fast traffic. Highways England's requirement was for machines that could lay and collect at least 400 cones at a rate of one every 10 seconds, cutting the crew from two workers to a driver. If you run traffic management, ask suppliers how their system handles the full range of cone types, lamps and signs in your layouts, not only the cones themselves.

Robot work is still street works

A robot on the carriageway does not change the rules that protect people around it. Safety at Street Works and Road Works: A Code of Practice, published by the Department for Transport, covers the signing, lighting and guarding of works on roads other than motorways and dual carriageways with speed limits of 50 mph or more. A crack-sealing machine or survey robot working in a live lane still needs a compliant site layout, competent operatives and a plan for pedestrians and cyclists. Ask suppliers to show how their machine fits standard layouts, where its operator stands and how it is protected from passing vehicles while it works.

Running a highway robot trial

Pick a stretch of network that represents the roads you maintain, with a mix of surface types and traffic levels, and collect your own baseline from recent inspection and repair records. During the trial, have inspectors check a sample of the robot's findings so you can count both missed defects and false alarms. Track repairs through at least one winter before judging their durability. Agree in advance who owns the survey data and how it will feed your asset management system. Tell residents and councillors what the machine is doing, because an unfamiliar robot on a street draws questions. Check insurance, vandalism risk and storage before the first shift, not after.

Sources and further reading

• Hertfordshire County Council: AI road maintenance robot — Local Government Association.
  Case study of a council's public road trial of a robot that finds road cracks and seals them before potholes form.
• Automated cone laying vehicle spotted on roads in new trial — Highways England, GOV.UK.
  2020 trial of machines that lay and collect traffic cones, with the safety reasons and the performance requirement.
• Safety at street works and road works: a code of practice — Department for Transport, GOV.UK.
  The code of practice on signing, lighting and guarding street works and road works on public highways.

Common questions

Can robots repair potholes without a crew?

Machines that find and seal cracks autonomously are being trialled on public roads, but most repairs still rely on crews and conventional plant. Treat autonomy claims as early-stage, ask for road trial results and plan for staff to supervise and protect any machine working on the highway.

Can AI spot potholes from ordinary survey footage?

Detection is a natural starting point for road robotics, because it does not need to touch the road. Before relying on it, compare its findings with your inspectors' records on the same streets and check how it copes with wet surfaces, shadows and parked cars.