clock menu more-arrow no yes mobile

Filed under:

Xbox One's new Kinect will likely hasten the pace of AI learning, says robotics scientist

The new Kinect included with the Xbox One could serve as a significant leap forward as computerized eyes for robots, said Cornell University computer science professor Ashutosh Saxena in an interview with Network World.

Saxena originally discovered the power of the original Kinect for robots applications when he brought home an Xbox 360 bundled with a Kinect sensor. He began using it in his robotics lab to allow robots to track human motions and actions; coupled with his software, Kinect helps the robots identify around 120 ordinary human activities such as eating, drinking, cleaning objects and taking medicine.

The existing Kinect's 3D depth-tracking capabilities are more advanced than the 2D tracking techology Saxena was using previously, and the device is much cheaper than the 2D equipment, too. According to Saxena, Kinect's inadequacies limit its potential applications, and it appears that the new Kinect won't suffer from those limitations.

"Modeling [hands] is extremely hard," he explained, pointing out that Kinect 2.0 can track individual fingers instead of just a hand.

The learning robots Saxena is working on could provide vital services for people with certain medical conditions, serving as a mechanical assistant to perform tasks like lifting objects and dispensing medication. According to Saxena, writing the software to go along with the hardware is the hard part, but the advancements that the new Kinect will offer could accelerate the timetable for the robots' commercial readiness — perhaps to a point not farther off than within the next five years.

While the Xbox One and its second-generation Kinect are set for release later this year, the Windows PC version of the new Kinect won't be available until 2014.

Sign up for the newsletter Sign up for Patch Notes

A weekly roundup of the best things from Polygon