The Telegraph: UAS that Can Dodge Obstacles Developed by US Scientists
An unmanned aircraft that can avoid obstacles without human input has been developed by US scientists.
Until now, the flying robots have always needed a human at the controls to operate them remotely and prevent them from crashing.
Now, however, researchers at New York’s Cornell University have managed to develop software that will help drones to dodge obstacles.
In the experiments, a miniature helicopter equipped with a camera captures images as it flies. Software developed by assistant professor of computer science Ashutosh Saxena and his team turns the image in the drone’s camera into a 3D model of its environment, and the robotic brain then uses an algorithm to determine which objects are obstacles.
In 53 flights in environments full of hazards, the robot managed to find a pathway without crashing into any obstacles – although the final two flights failed due to gusts of wind.
Eventually it is hoped that the drone will be able to calculate wind patterns and avoid moving objects such as birds.
The research is being funded by US Defence agency DARPA and drone-producer Lockheed Martin.
A paper entitled Low-Power Parallel Algorithms for Single Image based Obstacle Avoidance in Aerial Robots, outlines the success of the robot in experiments.
It says: “In outdoor robotic experiments, our algorithm was able to consistently produce clean, accurate obstacle maps which allowed our robot to avoid a wide variety of obstacles, including trees, poles and fences.”
To read the story on The Telegraph click here.