Drones with advanced algorithms could perform military tasks too dangerous for humans by autonomously scouting unknown environments and bringing back reconnaissance information.
The Defense Advanced Research Projects Agency (DARPA) in the US has completed the next phase of its drone program, where it demonstrated the capability of its advanced algorithms.
During its Phase 2 flight tests in a mock town in Georgia, the manufacturer says its drones showed they could:
Fly at increased speeds between multi-story buildings and through tight alleyways while identifying objects of interest;
Travel through a narrow window into a building and down a hallway searching rooms and creating a 3-D map of the interior; and
Identify and fly down a flight of stairs and exit the building through an open doorway.
During the flight tests, a team of engineers from the Massachusetts Institute of Technology (MIT) and Draper Laboratory reduced the number of onboard sensors to lighten their drone for higher speed.
J.C. Ledé, program manager, explained that the drone does not need to communicate with its launch vehicle so the chances of an adversary detecting troop presence based on radio transmissions are reduced.
A key part of the drone’s capability is to build not only a geographically accurate map as it traverses the cityscape but also a semantic one.
The drone’s onboard computer recognizes roads, buildings, cars, and other objects and identifies them on the map, providing clickable images as well which can be downloaded after the mission is completed.
Jon How, co-leader of the MIT/Draper team, said:
"As the vehicle uses its sensors to quickly explore and navigate obstacles in unknown environments, it is continually creating a map as it explores and remembers any place it has already been so it can return to the starting point by itself."
The MIT/Draper team has also made it possible to sync data collected by the drone with a handheld app called the Android Tactical Assault Kit, which has already been used by the British Army and US forces.
A separate team of researchers from the University of Pennsylvania (UPenn) reduced their own drone’s size and weight so it can fly autonomously in small, cluttered indoor spaces.
The UPenn drone can create a detailed 3-D map of unknown indoor spaces, avoid obstacles and fly down stairwells.
Camillo J. Taylor from UPenn said: "That’s very important in indoor environments because you need to actually not just reason about a slice of the world, you need to reason about what’s above you, what’s below you.
"You might need to fly around a table or a chair, so we’re forced to build a complete three-dimensional representation."
The next step, according to Taylor, is packing even more computation onto smaller platforms, potentially making a smart UAV for troops or first responders that is small enough to fit in the palm of the hand.