Autonomous drone maps 3D models of dense urban environments

January 22, 2020 //By Rich Pell
Autonomous drone maps 3D models of dense urban environments
Defense contractor and industrial corporation Raytheon (Waltham, MA) and aerial robotics company Exyn Technologies (Philadelphia, PA) have co-developed a mapping autonomous drone (MAD) - a fully autonomous aerial robot that uses artificial intelligence (AI) and can operate in GPS-denied environments to map dense urban environments in 3D.

The drone, which can travel deep to reveal tunnels, urban undergrounds, and natural cave networks, has been demonstrated to the U.S. Department of Defense, says Raytheon.

"Operators will have to do very little, because most of the reasoning is done on board with artificial intelligence," says Mark Bigham, chief innovation officer at Raytheon's Intelligence, Information and Services business. "They simply set the boundaries for the robot to explore and press 'go.'"

While most drones use GPS to navigate, the MAD is able to approximate its position on the map while flying through GPS-denied environments such as fortified buildings and underground facilities. Using its own estimate, says the company, it can maneuver over the terrain without GPS, indoors and out.

The drone uses a variety of sensors to gather 300,000 data points per second, mapping as it moves along its pathway. The system uses lidar - which measures distance to a target by illuminating the target with infrared laser light and measuring the reflected light with a sensor - to navigate and build a 3-D map.

"The drone detects obstructions and obstacles like rubble and rebar, dangling wires, then methodically works around them," says Bigham.

Raytheon installed its proprietary technology to georegister maps and objects where GPS is not available - a feature that can be critical to the success of certain missions, says the company.

"Forces and first responders are going to need to know where things are down to the centimeter," says Bigham.

MAD, says the company, has achieved three "firsts":

  • the first time an autonomous drone has gone from mapping the outside of a building to the inside;
  • the first floor-by-floor mapping of a multi-level building;
  • and the first time machine learning has been used to locate objects in a three-dimensional space, all in real time.

"Terrorists and guerillas have literally gone underground to hide from satellites, larger drones, and patrols," says Bigham. "This system can help us


Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.