In 2001, a group of researchers at Vanderbilt teamed with NASA in a project designed to create software aimed at building robotic machines which learn how to function in a similar way to humans.
The project took on Vanderbilt’s Center for Technology Transfer and Commercialization as a partner to bring the work to market, and seven years later the results of the project were realized.
The company, known as Universal Robotics, produced a technology called “Neocortex,” a form of software that mimics human learning in order to adapt and work well in unfamiliar or chaotic environments.
Now, 12 years after the project was started, the International Space Station plans to deploy the technology to perform maintenance tasks which are difficult for astronauts to accomplish.
According to CEO David Peters and vice president Hob Wubbena, Neocortex works by ignoring distracting and impertinent information and concentrating the machine’s attention on useful sensory details. The machine distinguishes between what works and what fails, and in this sense learns not to perform tasks in the way which failed, increasing the success rate of the machine significantly.
This is reportedly the first time any such technology has been put to use, let alone attempted or created.
“There’s nobody in the world that has anything close to this,” said Wubbena proudly.
Neocortex can be installed on any type of automated machine or robot, giving the technology huge prospects for the future.
Situations and tasks which involve a high rate of worker industry are where the company is focused now, including any commercial work involving the handling of materials or logistics.
As Peters said, after countless years of research and testing, “We’ve gotten pretty good at predicting the unknown,” which opens many doors for possible applications of Neocortex down the road.
Edited by Brooke Neuman