clock menu more-arrow no yes mobile

Filed under:

MIT researchers program autonomous airplane to navigate in tight spaces without GPS

New, 9 comments

Researchers at MIT have created an algorithm to pilot an autonomous airplane without using GPS, a significant challenge because of the limitations of fixed-wing aircraft.

MIT autonomous plane
MIT autonomous plane

Most unmanned drones designed for precision flying are helicopter-like, from hexacopters created for aerial filming, to the University of Pennsylvania's adorable synchronized quadrocopters. Now, researchers at MIT have created an algorithm to pilot an autonomous airplane without using GPS, a much bigger challenge because of the limitations that come with fixed-wing aircraft: no hovering, no sideways movement, and the need to maintain enough speed to keep it in the air. The MIT team designed a plane with short and broad wings specifically for the challenge, which allow it to fly relatively slowly, make fairly precise turns, and support the weight of the necessary electronics.

The plane uses a combination of two algorithms to determine its location, orientation, velocity, and acceleration, which requires the craft to calculate 15 different values. Because this is such a massive computational challenge, researchers gave their plane the benefit of an accurate digital map of its environment, but the project's next step is developing an algorithm to build such a map while in flight. The fact that the drone operates without the help of GPS is also significant, and such advances could eventually be applied to prevent the capture of drones via GPS-spoofing. While still a very new area of research, MIT's early successes and yet-to-be-overcome obstacles are paving the way for the future of autonomous plane navigation and GPS-free drone navigation.