Insight

AI Special: Making construction more productive

31 January 2018

New York company Construction Robotics created the first commercial bricklaying robot, called SAM, or semi-automated mason

Machines that can learn tasks will be a familiar sight on site, says Denise Chevin.

The addition of artificial intelligence and machine learning to robots and image recognition software will be increasingly harnessed to improve processes and productivity.

Graphics hardware giant Nvidia has teamed up with a number of construction firms to supply its chip with AI capabilities. One is Komatsu, Japan’s largest maker of construction machinery, which is aiming to create a generation of products with embedded AI, as a step towards unmanned equipment.

Nvidia will also supply AI chips that allow Komatsu machinery to “see” the site and react to developments as they take place. They will be complemented by drones equipped with similar graphics processors.

AI Special: Related articles

Rian Whitton, an analyst with consultant ABI Research, highlights the potential of smart drones to  affect surveying, monitoring and inspection in construction.

“For the most part, UAVs collect data from a site, upload it to a cloud, after which bespoke software garners insights from the data. Having on-device machine-learning capability could provide that service in real time, allowing managers and operators to understand and react immediately to developments on the ground.”

He adds: “Expand that to multiple platforms that fly, land, dock and recharge autonomously, and you have a much more comprehensive understanding of the project.”

This will allow better allocation of resources, and help identify efficient routes to move material through the site, he adds. Beyond maintenance, having real-time analysis of a site will help contractors at the surveying, estimating and planning phases.

So how far are we down this road? “Not very,” says Whitton. “Commercial drone opportunities have only begun to gain momentum in construction, and with them comes a range of AI/ML solutions. I would suggest some of the technologies being talked about will reach maturity within three years.”

The development of robots equipped with machine learning is seen as having huge potential. Where traditional robots had to be programmed to do a limited number of tasks in a specified environment, the new generation is not as rigid, says Autodesk’s Tatjana Dzambazova.

Training the robot

“In our robotic lab in San Francisco, we are currently piloting a construction site robot that can install glass facade panels from the interior of a building, a task that is usually quite laborious and dangerous, especially when installing on upper floors,” she explains. “We initially train the robot by having a human guide its movements via a virtual reality set. Through this training the robot learns the nature of the task to find a glass panel in a certain area, to lift it in a certain way and place it in a recognisable location.”

Smart robots are springing up to do tasks from installing rebar (see p33) to bricklaying. One, called Hadrian, can lay 1,000 bricks an hour, day and night, and could theoretically build 150 homes a year. But there is also debate as to where their future lies.

Nick Leach, strategic BIM manager with Sir Robert McAlpine, thinks it may take a generation before robotics on site are an industry norm.

“The real immediate potential I see taking an effect on construction in the short term is around the automation with computing/machine learning – where utilisation of data to analyse not just design but site-based tasks and management activities will improve how projects are delivered, raising the bar further in the quality and safety of our construction sites.” He cites image recognition systems with multiple uses as an example of technology that will gain momentum.

“I would say it’s an exciting time. Companies are starting to realise the potential, but with developments happening on so many fronts it can be a minefield for organisations in deciding on the right way to go.”

Graphisoft’s Ákos Pfemeter also believes that the future is more about developing smart factory-based machines. “The size of bricks has not changed for 500 years; they are optimised for human labour. It doesn’t make any sense to introduce a machine to replace directly how a human would do it,” he says.

Paul Cook, head of technology at ISG Technology Solutions, thinks AI will continue to refine the technology used. “But I don’t believe these advancements will lead to huge strides in the design and performance of buildings and the construction process, unless we fundamentally change the very design approach that is used – a shift from designing and procuring in silos to implementation via a master systems architect.

“Without this radical transition, technology such as AI, VR and AR will not reach its optimum potential, because the platforms and ecosystems they use are not seamlessly integrated.”

Leave a comment