How McLaren Construction is deploying autonomous robot dogs on sites

The contractor will focus deployment, at least initially, on sites where the technology will have the most impact. Thomas Flannery, head of digital information management at McLaren, tells DC+ that the deployment will embrace the variety of sectors McLaren works in. “We do data centres, arenas, commercial, fit-out, residential. We want to understand which sectors the technology will bring the most value to,” he says.

Naturally, for a project to feature in the initial deployment, it must have “robust approaches to digital construction and good-quality models coming through the design team”, he notes. McLaren will also liaise with its clients. “A lot of our clients are very keen to see what we’re doing in this space, and they’re very open and receptive to seeing this technology on their sites,” Flannery says.

Gothard explains how a typical deployment will begin: “FieldAI will train the dog on the site environment; they’ll input data, such as the drawings and models, so it can navigate by itself. [A few months later] we’ll take over the set-up ourselves with some light-touch management from FieldAI.

“The robot will conduct mapping walks under manual control. It starts to collect data – LiDAR, point cloud, photogrammetry – and starts to understand its environment in its current context. Then we feed it the drawing as well, so it has something to relate that in terms of a floorplan. The models are helpful, but not necessary.”

Once the robot ‘knows’ the site, its mission and schedule are set, including “what it’s going to capture and any exclusion zones – such as a concrete pour”, he adds.

The logistics will need to be factored in. Gothard explains: “Say our project’s got a floor area of 30,000 sq m: how many walks is that and how does that affect the dog’s battery life? How often does it need to go back to its docking station, charge up, and then redeploy? What time will the robot do its walk: before starts, at lunchtime, or after work finishes?

“We would like the dog to conduct daily walkarounds for general 360-degree capture, just so we’re never missing anything, any part of the installation or any step in the construction of our projects. Capturing point cloud data will probably be at a lower cadence; otherwise we’ll be overprocessing the data: there’ll be too much to handle and make sense of.”

Flannery adds: “The dog will walk the site, do the point cloud capture, and go back to its docking station. Then the point cloud data will [be fed] to the Naska module. With the models already there, Naska will do the comparison. It takes a day, maybe, to turn it around. Now we’re discussing with Naska about getting the outputs and getting those straight to who needs to see them – obviously the quality team and the site manager – so they can let the subcontractors know what needs to be fixed before they have moved on from that area.”

Flannery describes McLaren’s move into autonomous robots as a three-year commitment. “If the primary use cases [the photogrammetry to OpenSpace and the point cloud scans to Naska for deviation analysis] are satisfied, we’ll see a return on investment. Anything else over and above that is a value-add,” he says.

Gothard cites the cost savings of having the dog carry out the photogrammetry on a project – rather than site managers – at £30,000 to £35,000 a year. The added benefit is the site manager having their time freed up to concentrate on “what humans are better at – which is talking to people, having those conversations, looking out for certain items, whether it be health and safety, quality, environmental, those sorts of bits on site”.

The FieldAI dogs also represent a significant saving on laser scan surveys.

But it’s not just about the primary use cases: McLaren’s digital team is alert to other opportunities that may arise. Gothard adds: “We’re really just getting started on full deployment onto projects now. I think this is the stage where we’re really going to get the most lessons in terms of building out the ROI case, understanding what the technology’s full capabilities are. We’re challenging our project teams to come up with practical ideas [of what else the FieldAI dogs can do]: we’re asking what they’re trying to track in their project, what their key construction risks are.”

Flannery notes: “I’m not going to name it, but on one project, the team’s got a real issue with intumescent paint on the steelwork and tracking damage. We’ve developed a use case with FieldAI to get the dog out scanning once the steel’s installed, and then feed that back, getting the data as quickly as possible to understand where the damage is on site, because it’s quite a lot of steel.”

Of course, McLaren’s move isn’t just about improved efficiency and quality for the sake of its own bottom line. “We’re getting a lot of projects now that require us to validate every piece of installation,” Gothard states.

McLaren is planning to trial a bricklaying robot, “not to replace humans, but to meet skill shortages”, says Gothard.

The contractor’s digital team has also recently coordinated a group approach to deploying drones. “Obviously, there’s quite a developed regulatory framework around deploying drones, so we have to follow quite a specific process for that, and I’m sure robotics will probably catch up to that,” Gothard notes

Flannery emphasises that these developments are part of McLaren’s ongoing digital transformation. “It’s taken four years of work. Our investment plan since 2022 has seen us rationalise our technology stack and invest in individuals like Tom – he’s not only the implementation manager for FieldAI; he also brought Dalux on to our sites as well. Our data stack is now in a position where we can start leveraging AI across our project portfolio. That investment has got us to this position now where we’re confident that deploying AI-enabled autonomous robots on site will be a success.”

Keep up to date with DC+: sign up for the midweek newsletter.