Image: DARPA
DARPA’s Fast Lightweight Autonomy (FLA) program is betting big on the autonomy of drones. During a recent test held in a messy warehouse in Massachusetts, their sensor-equipped quadcopter achieved a target speed of 45 miles an hour while dodging numerous objects in real-time. And it did this autonomously.
The FLA program is working towards the development of algorithms that could reduce the load of power, communications, and human intervention need for drones to achieve low-level tasks (navigating obstacles in a messy environment). If deemed successful, these algorithms could reduce the workload of human operators on simple tasks allowing more human supervision on higher-level tasks.
Picture this: a military group or rescue team is patrolling a dangerous urban environment such as a collapsed building or hostile-ridden compound. Today’s drones could provide aerial imaging for those two scenarios, but to provide a view on the inside would require humans to actually go in. FLA algorithms could, instead, give drones the ability to scout the quandary that is indoor environments autonomously leaving human lives outside the equation.
The drone, made with DJI’s Flamewheel airframe, 12-inch propellers, and a few sensors, successfully demonstrated the algorithm’s use by zooming though a cluttered warehouse at 45 miles an hour unmanned. Sensors onboard include high-definition cameras, LIDAR, sonar, and inertial measurement units. Dodging walls and obviously-marked boxes was achievable at low-speeds, but DARPA hopes to improve the algorithm and computing speed to where the drone’s mass is taken into account allowing it to make high-speed maneuvers.
The team at DARPA is hoping that their technology could guide drones though windows (and similar size-restricting areas) without operators and GPS.
Source: DARPA
The post DARPA Drone Dodges Objects At 45 MPH, Autonomously appeared first on SimpleBotics - Covering The Evolving World Of Robotics And Drones.