Boston-based GE Subsidiary Avitas Systems is using drones, wheeled robots and autonomous underwater vehicles to collect images required for inspections in the oil, gas, energy and transportation industries, and it’s all powered with Nvidia’s purpose-built AI supercomputers.
The partnership marries Nvidia’s DGX-1 and DGX Station AI computers with the inspection services of Avitas Systems using Nvidia’s off-the-shelf machine-learning technology.
And even though companies like Kespry and Flyability have been using drones for monitoring mines, inspecting wind turbines, and assessing building insurance claims for some time, the rapid advances being made in machine learning, coupled with low-cost drones and robotic systems is making it possible to automate whole sectors of low-skill work.
Alex Tepper, the founder and head of corporate and business development at Avitas said, “We are using artificial intelligence to do what we call automated defect recognition, which interprets the sensor information from the robots to detect defects of various kinds – like corrosion, hot spots, cold spots, and micro-fractures.”
The two Nvidia models of the DGX-1 supercomputers will perform different functions, with one performing at 960 teraflops and another that reaches 170 teraflops, according to a Nvidia spokesperson. The DGX-1 is housed in a data center and includes a complete software stack and an integrated operating system, middleware components, and the software tools. This allows Avitas developers to write their algorithms.
Avitas is also using the DGX Station, a desktop supercomputer that packs the power of half a row of servers in a neat package. This makes it easier for Avitas to power their automated systems in n remote areas, where it’s hard to get good network connections, such as oil rigs in the middle of the ocean, or facilities in the desert.
“It’s too expensive to do processing of data and moving it back and forth, so by having the equivalent of half a row of servers out in the field is the advantage of it, said Jim McHugh, general manager of DGX Systems for Nvidia.
“Using our latest DGX systems to help train robots and better predict industrial defects increases worker safety, protects the environment, and leads to substantial cost savings for companies,” McHugh added.
The whole point is that with automation and machine-learning, inspection costs can be reduced by 25 percent. As an example, a mid-sized refinery that has been spending $4.0 million a year on inspections can reduce the cost to $3.0 million using Avitas’ services.
Advances in artificial intelligence have made it possible to teach robots to navigate by themselves, learning to avoid obstacles. Just this week, Neurala, a company that specializes in deep learning, launched a drone toolkit, the Brains4Bots that can be used to train a vehicle to recognize or follow a particular object and avoid obstacles. The Neurala Brain is based on technology originally developed for NASA and the U.S. Air Force
