MIT’s Computer Science and Artificial Intelligence Lab researcher developed an obstacle-detection system that allows drones to avoid autonomously obstacles at 30 miles per hour in a tree-filled field.
The student-researcher Andrew Barry, who developed the system with MIT professor Russ Tedrake as part of his thesis, says “Everyone is building drones these days, but nobody knows how to get them to stop running into things” He says that sensors like Lidar, which are too heavy to put on a small aircraft aren’t practical and that also the system of creating maps of the environment in advance is also not practical.
- MIT researcher Barry says “If we want drones that can fly quickly and navigate in the real world, we need better, faster algorithms,”
This system runs 20 times faster than any existing software. The student’s stereovision algorithm allows the drone to detect objects and build a full map of its surroundings in real-time. The software is operating and 120 frames per second, it is open-source and available online.