Getting drones up in the air isn’t the problem; it’s what happens to the drone once it is in the air that’s the problem. If you aren’t a seasoned drone operator, you may struggle to get your drones to fly around objects without hitting them, and you may even lose your drone to an accident. Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) are trying to address this problem.

The researchers have created motion-planning algorithms that segment space into “obstacle-free regions” in order to find a collision-free route. The algorithms have been previously used to plan footsteps for CSAIL’s Atlas robot.

“Rather than plan paths based on the number of obstacles in the environment, it’s much more manageable to look at the inverse: the segments of space that are ‘free’ for the drone to travel through,” said recent graduate Benoit Landry. “Using free-space segments is a more ‘glass-half-full’ approach that works far better for drones in small, cluttered spaces.”

In a video demonstrating the use of the algorithms, the researchers show a small quadrotor doing donuts and figure-eights through an obstacle course. In another video, the team demonstrates a drone avoiding obstacles without any previous knowledge of the space.

“As the drone flies, it continuously searches through the library to stitch together a series of paths that are computationally guaranteed to avoid obstacles,” said Anirudha Majumdar, a Ph.D. student at MIT. “Many of the individual funnels will not be collision-free, but with a large-enough library, you can be certain that your route will be clear.”

The drones aren’t able to avoid obstacles at high speed yet, but the algorithms do show they are capable of navigating through complex, smaller and denser environments. The researchers hope their work will continue to be built upon and improved.

“A big challenge for industry is determining which technologies are actually mature enough to use in real products,” Landry said. “The best way to do that is to conduct experiments that focus on all of the corner cases and can demonstrate that algorithms like these will actually work 99.999% of the time.”