Air Force

UAV recovery from a moving platform

Sophisticated and robust image recognition software integrated with flight controls steers a drone to a safe recovery

Software & Information Technology

Commercial fishing vessels utilize UAVs to scan for schools of fish, and this is just one of the many emerging uses for the burgeoning field of drone applications. Launching the UAV from a deck mounted catapult is a relatively straightforward mechanical process. Retrieving one is not.

This Air Force invention offers an improved method for landing a UAV on a ship or other moving vehicle.

Capturing a UAV at the end of a flight requires an operator to use a radio transmitter to send flight control commands to the UAV to fly it into a recovery device, such as a recovery net, or arresting cable device. Even with the aid of guidance methods, such as Global Position System (GPS) or Inertial Measurement Unit (IMU) sensor-based navigation, skilled UAV operators require extensive training and experience to safely recover these expensive drones. Onboard a surface in motion such as a truck or ship, successful recovery is challenging.

The image processing software is a key to the system and is unique as a result of its simplicity and high efficiency. This allows it to operate very quickly and robustly to identify the correct target on the net and to do so fast enough to allow the UAV to correct its flight path in time to make a recovery.

Conceptual illustration of the UAV recovery system.

As an operator of many UAVs, it has been necessary for the Air Force to develop a method for drone recovery from a moving platform. In this recently developed system, a UAV is guided with GPS to within approximately 100 m of the ship or vehicle. Meanwhile, the UAV’s onboard video camera captures images which are sent to the surface control station.

When the object tracking program is executed, this live feed is displayed at the control station. Image tracking algorithms allow the UAV to identify a unique set of patterns and colors on a retrieval net or other recovery surface. The algorithm finds the object which best matches the parameters selected by the user and calculates the coordinates of that object’s center. These are transformed into the aircraft’s coordinate system, and range, bearing, and azimuth are sent to the UAV by the control station. The UAV then uses its onboard autopilot to home in on the target. Coordinates are generated for each video frame and continuously sent to the UAV, and the feedback loop ends when the user terminates the program.

Do you have questions or need more information on a specific technology? Let's talk.

Contact Us