In March 2011, the Tohoku earthquake in Japan initiated a tsunami that wreaked havoc across the coast of the island nation. The most infamous damage occurred when the Fukushima Daichii nuclear power plant suffered serious equipment failures, which led to the second largest nuclear disaster in history.
In the aftermath, inspection and repair crews faced severe health threats as they navigated the radiation-ridden facility. But what if humans didn’t have to clean up the wreckage, asked Jerry Hajjar, professor and chair of the Department of Civil and Environmental Engineering at Northeastern. “What if you could imagine an autonomous flying vehicle or a swarm of vehicles that would be able to go in instead.”
Hajjar knows that any major operation that’s difficult to access because it’s too remote, too dangerous, or too large could benefit from a robotic “inspection assistant.” And now he’s teaming up with researchers at Carnegie Mellon University to create one.
Backed by a $2.3 million grant from the National Science Foundation’s National Robotics Initiative, the research team includes robotics and image-processing experts who will develop an autonomous low-flying vehicle equipped with laser and video capture technologies capable of providing nearly continuous data about a given structure.
Hajjar, for his part, will develop algorithms that automatically interpret the collected data. “Given what we’re seeing, I’ll try to determine and validate whether or not the structure we’re looking at is damaged, how it’s damaged, and possibly how it got damaged,” he explained.
He will be linking existing structural analysis models with image data. “This is the first time we really have a ubiquitous set of data on the structure,” he said. In the past, models have relied on sensors that scan for quantities such as strain or stress at a single point. Here, through the image processing expertise of his colleagues, Hajjar will be working with information about the deformations exerted across the entire structure, be it a failed power plant, an aging suspension bridge, or a collapsed building.