From bridges to oil refineries and defense, robots fix on aging U.S. infrastructure

The infrastructure that makes up the U.S. is, in some cases, flailing. More than a third (36%) of all U.S. bridges — nearly 222,000 spans — require major repair work or replacement, according to Department of Transportation data. Meanwhile, more than 1,500 people were injured or died at 153 oil refineries across the country in the six years ending in 2023, in part because of sites unwittingly operating damaged structures. In addition to the risk to human safety, damage to an oil refinery causes an increase in the already strong pollution factor. And with climate shocks changing the world and its structures at an increasingly rapid rate, it can be difficult for even the most experienced workers to keep up. That’s where robots come in.

“What’s the process of trying to evaluate physical structures like bridges or dams or ships? It’s a very archaic method of, put a human up there […] and write down all the potential damage areas there are, which doesn’t work that well when you’ve got miles and miles of potential damage areas,” said Jake Loosararian, CEO and co-founder of Gecko Robotics.

Gecko deploys robots that climb, crawl, swim and fly, which work with sensors and artificial intelligence software to create layers of data that tell you all there is to know about a structure’s health. It can do all of this without putting humans in the line of fire, even though they are the experts who pilot the robots, then make sense of and act on the tech’s findings.

Gecko, which launched out of Loosararian’s college dorm in 2013 after he visited a Western Pennsylvania power plant and learned about its failures and explosions, isn’t the only one tackling the infrastructure problem with robotics. Percepto offers an end-to-end drone-in-a-box solution that is equipped with different types of cameras and pairs with its proprietary inspection and monitoring software.

“If you’re talking about the oil and gas industry, you can find gas leaks, you can find crude oil leaks, and you can find broken components,” said Udi Zohar, Percepto’s chief product officer. Zohar added that its drone hardware can get to electrical infrastructure in inclement weather conditions that humans cannot (like hurricanes or snowstorms, for example).

The drone software also receives alerts and dispatches drones to determine the accuracy of the alert. “You’re saving people time. You’re adding safety to your company,” said Zohar. While Percepto started out with the main use case of providing security at sensitive facilities, it has expanded to offer a range of use cases, including pinpointing weak spots where methane emits or getting utilities up and running after climate disasters.

At Gecko, the data is just as important as the robots.

“Initially, we built wall-climbing robots that could basically do sonograms just like a doctor does for pregnancy scans, but for a physical structure,” said Loosararian. By attaching ultrasonic sensors and cameras, robots could not only access difficult-to-reach and often dangerous spots, but also be able to capture information and evaluate changes over time, as well as the root causes of that change. “That begins to affect how you run certain things, or how you predict how long something has to last, or how safe something is,” he said.

Companies like Gecko and Percepto are targeting use cases that increase safety and efficiency, and reduce expenditures, in industries like defense, energy and manufacturing. Operation and maintenance (including that of aging assets) accounts for 39% of the U.S. military budget, and the national total deferred maintenance cost may be at least $1 trillion, research shows. By employing robots to proactively tackle structural assessment and prioritize repairs by need, companies and agencies can create a healthier, safer built world, often for less money.

Loosararian said the infrastructure-inspecting robotics industry is growing so fast he can’t take it on alone. “My problem is I need more robots, and I can’t build them all,” he said.

Meanwhile, Gecko is actively working on an AI model that predicts when assets will fail. Percepto, for its part, is evolving to include more types of cameras and sensors on the drone-in-a-box product. These two companies reflect a greater trajectory — the inspection robot industry at is growing at a rate of at least 30% per year.

In the U.S., motorists cross bridges rated in poor condition roughly 168 million times a day. Hundreds of workers inspecting sites fall to their deaths each year. And at oil refineries, factors like corrosion increase dangerous emissions that harm nearby workers and the planet at large, while simultaneously increasing operational costs.

While much-needed experts continue to seek structural anomalies, robots — and the people operating them — are swooping in to fill a clear gap for a safer, healthier world.