Drones crashing is nothing new. New drones are becoming more sophisticated in this area though. PhD student Andrew Barry from MIT’s Computer Science and Artificial Intelligence Lab has come up with a obstacle-detection system that allows drones to autonomously fly through trees at 30 miles per hour.
The system relies on a stereo-vision algorithm that allows drones to detect objects and build a full map of their surroundings in real-time. The software is open-source and operates at 120 frames per second. More information is available here.