MIT has developed yet another piece of technology best described as “awesome yet vaguely terrifying should the Machines ever rise up against us.”
The school’s Computer Science and Artificial Intelligence Laboratory (CSAIL) has unveiled software capable of piloting drones with uncanny precision, using complex motion-planning algorithms. The lab released the above video of one of their quadrotors bobbing and weaving through a simulated “forest” of PVC pipe and tripwires, a la Mission Impossible. The little guy, capable of flying through a 10-square-foot space at speeds upwards of a meter per second, weighs just over an ounce and measures 3.5 inches from rotor to rotor.
The drone uses algorithms, previously used to plan footsteps for CSAIL’s Atlas robot, to detect “obstacle-free regions” in space, a technique researcher Benoit Landry calls “Iterative Regional Inflation by semidefinite programming,” or IRIS. This is combined with a “mixed-integer semidefinite program,” which plots the drone’s course along a single, uninterrupted route in 10 minutes or so.
“Rather than plan paths based on the number of obstacles in the environment, it’s much more manageable to look at the inverse: the segments of space that are ‘free’ for the drone to travel through,” Landry said in a release. “Using free-space segments is a more ‘glass-half-full’ approach that works far better for drones in small, cluttered spaces.”
To get its bearings, the drone uses motion-capture optical sensors and an onboard inertial measurement unit. “I’m most impressed by the team’s ingenious technique of combining on- and off-board sensors to determine the drone’s location,” said Jingjin Yu, assistant professor of computer science at Rutgers University. “This is key to the system’s ability to create unique routes for each set of obstacles.
The researchers at CSAIL have made their motion-planning algorithms available to anybody. You can find them here.
Source URL: https://www.bostonmagazine.com/news/2016/01/20/mit-drones-video/
Copyright ©2020 Boston Magazine unless otherwise noted.