AI can help contractors track progress, spot dangerous behaviors and prevent accidents.
AI will one day completely transform the jobsite — construction equipment will leverage machine learning to become more and more adept at performing complex tasks autonomously. In the meantime, though, here are three ways AI is changing construction now, at least for large contractors with deep pockets.
Tracking project progress
Nothing throws off a schedule like a construction deficiency that’s discovered when the work is almost completed. AI systems can help contractors keep a closer eye on all parts of a project throughout the construction process so they can make any necessary corrections right away.
An AI system from startup company Doxel uses drones and autonomous land-based vehicles to take daily photos and laser scans. A deep neural network that’s been taught to recognize construction site features then analyzes the field data and compares it to project models to identify any deviations so errors can be corrected immediately. The AI software also measures installed quantities and lets contractors compare work completed to project schedules so they can tell if they’re where they need to be to deliver a project on time.
Drones have made it easy to get overview images of jobsites. Add AI into the mix and you end up with a whole new way of visualizing and analyzing jobsites and tracking materials, people and conditions.
Japanese equipment manufacturer Komatsu recently partnered with tech company NVIDIA, known for PC video games, to bring AI to jobsites. NVIDIA’s graphical processing units, or GPUs, will communicate with drones and cameras on the jobsite, and its AI platform will use the data to create 360-degree 3D visualizations of the site.
Stereo cameras installed in the cabs of heavy equipment will work with the AI system to report on changing jobsite conditions, showing in real time the positions of people, machinery and objects. That should reduce collisions and other accidents.
Spotting dangerous behavior
No matter how many times you emphasize safety in toolbox talks and other training, some workers neglect basic precautions, including wearing proper PPE. AI can help managers identify unsafe behaviors so they can intervene before an injury occurs.
Using artificial neural networks, a team of machine-learning researchers in China and Australia developed a model that can analyze images from the jobsite and identify workers who are not wearing safety harnesses when working at heights.
Smartvid.io has developed a similar system that can analyze video and images from a number of different sources including jobsite cameras, drones and workers’ phones. A safety manager can search through videos and photos using tags such as “hardhats” or “ladders” to help them identify specific problem behaviors.
AI won’t be able to solve every problem on the jobsite — construction will always require human brains — but it’s reaching the point where it knows enough to provide information and insights that can help improve project outcomes.
Freelance writer Mary Lou Jay writes about business and technical developments in a variety of industries. She has been covering residential and commercial construction for more than 25 years.