Cornell researchers have created an autonomous flying robot that they say can maneuver around obstacles as efficiently as a bird.
Developed with funding from DARPA, the researchers say it could be of great value in search-and-rescue operations thanks to its ability to guide itself through forests, tunnels or damaged buildings.
The robot’s based on a quadrotor, a commercially-available flying machine about the size of a card table with four helicopter rotors. The Cornell team’s already programmed quadrotors to navigate hallways and stairwells using 3D cameras, but has found that these cameras aren’t accurate enough at large distances to plan a route around obstacles.
To get round this, assistant professor of computer science Ashutosh Saxena is working to turn a flat video camera image into a 3D model of the environment using, for example, converging straight lines, the apparent size of familiar objects and what objects are in front of or behind each other – the same cues humans unconsciously use.
The robot has been trained with 3D pictures of such obstacles as tree branches, poles, fences and buildings, learning the characteristics all the images have in common, such as color, shape, texture and context. The resulting set of rules for identifying obstacles is burned into a chip before the robot flies.
In flight, the robot breaks the current 3D image of its environment into small chunks based on obvious boundaries, decides which ones are obstacles and computes a path through them as close as possible to the route it has been told to follow, constantly making adjustments as the view changes.
It’s been tested in 53 autonomous flights in obstacle-rich environments, failing only twice because of winds.
