DARPA

The US Defense Advanced Research Projects Agency (DARPA) has tested a self-navigating quadcopter using only onboard sensors / software at an old hangar set up as a warehouse at Otis Air National Guard Base in Massachusetts, US.

The test was part of DARPA’s fast lightweight autonomy (FLA) programme, which intends to develop and test algorithms that can reduce human intervention needed to fly small unmanned aerial vehicles (UAVs) around a crowded urban surrounding.

The FLA programme also aims to develop and demonstrate autonomous UAVs small enough to fit through an open window, flying at a speed of at 20m per second without any outside help or dependance on GPS.

During the recent flight test, researchers at DARPA were able to achieve the targeted speed of 20m per second for the unmanned quadcopters using a commercial DJI Flamewheel 450 airframe, E600 motors with 12in propellers and 3DR Pixhawk autopilot as the common quadcopter UAV platform for the test.

The test demonstrated the testing of high-definition onboard cameras and other sensors, including LIDAR, sonar and inertial measurement units, which use sound and light respectively to identify the location of surrounding objects.

The initial autonomous capability of the quadcopters, including ‘seeing’ an object, was also tested.

"We are excited that we were able to validate the airspeed goal during this first flight data collection."

Following successful completion of the programme, FLA technologies will be able to reduce operator workload and stress, and can be used by military teams in critical missions in urban environments and for rescue teams in disaster management operations.

The small UAVs, to be developed under the FLA programme, will be able to quickly navigate through a maze of rooms, stairways and corridors or other cluttered areas without a remote pilot.

DARPA programme manager Mark Micire said: "We are excited that we were able to validate the airspeed goal during this first flight data collection.

"The fact that some teams also demonstrated basic autonomous flight ahead of schedule was an added bonus."

Two years ago, DARPA started a challenge for researchers to build an autonomous drone. DARPA and three other research teams are trying to advance the onboard algorithm.

The teams that competed were the Massachusetts Institute of Technology collaborating with US-based nonprofit engineer laboratory Draper, the University of Pennsylvania and technology firm Scientific Systems in Woburn, Massachusetts, US in collaboration with AeroVironment.

DARPA did not name the best performer and released a video showing performances of the drones.

Image: A FLA quadcopter self-navigates around boxes during initial flight data collection. Photo: courtesy of DARPA.