Drones are no longer in the realms of geekdom and sci-fi movies. Companies like Amazon and McDonald are already considering the drones-delivering services in major cities of the world. Even here in India the police are using drones to monitor communally troubled areas where policemen themselves cannot enter due to security reasons. They can be remote controlled or self-driven (or rather, self-flown).
After this what remains is improving upon the current technologies. For example, in the video given below, you can see a self-flying drone that dodges through the trees in a jungle at 30 mph. It doesn’t just dodge trees it also avoids obstacles like stones and the branches of the trees. With the help of a camera, the drone quickly processes the information that it gathers while flying and then in real-time decides which turn it should take in order to avoid crashing into the object.
The video has been released by a research team from the Massachusetts Institute of Technology (MIT). It is claimed to be the most advanced system of object detection and avoidance seen to date.
According to CSAIL PhD student Andrew Barry who developed the system as part of his thesis with MIT professor Russ Tedrake, “Everyone is building drones these days, but nobody knows how to get them to stop running into things. Sensors like lidar are too heavy to put on small aircraft, and creating maps of the environment in advance isn’t practical. If we want drones that can fly quickly and navigate in the real world, we need better, faster algorithms.”
The drone, called Pushroom Stereo, moves 20 times faster than its contemporary versions using stereo-vision algorithm that allows the drone to detect objects and build a full map of its surroundings immediately. It operates at 120 frames per second. It can extract depth information at a speed of 8.3 ms per frame. The software that performs all these calculations is open source and if you would like to create something of yours, you can download it here.
The Pushroom Stereo drone has a wingspan of 34 inches and it weighs around 1 pound. It was assembled using off-the-shelf components costing around $1700 including a camera on each wing and two processors not more complicated than the one your mobile phone uses.
Here is the video of the self-flying drone that avoids obstacles like trees and branches and stones while flying at 30 mph.