BAE Systems is flying a modified Jetstream regional turboprop to test automatic flight control systems for unmanned air vehicles, which can fly - and in emergencies also land - UAVs without ground station input.
The objective is to provide UAVs with sufficient autonomy to avoid other aircraft and hazardous weather without using a ground-based pilot to make decisions using real-time video transmissions. The work is part of the UK-led Autonomous Systems Technology Related Airborne Evaluation and Assessment (ASTRAEA) programme to allow UAV flight in normal airspace.
The UK manufacturer has conducted about a dozen test flights with the Warton-based turboprop since last September, and plans about 25 more before the ASTRAEA programme ends in March 2013. While take-offs and landings are manually flown by two onboard pilots - these manoeuvres are relatively simple and pose no automation challenge, says Darren Ansell, engineering manager - the programme focuses on the intermediate flight phases.
ADS-B positioning data from other aircraft is used to form a picture of the local traffic situation. This information may, however, not be available from some airspace users, such as light aircraft, hot air balloons and parachutists, says Lambert Dopping-Hepenstal, BAE's engineering director systems and strategy.
As a result, the Jetstream has been equipped with a video camera in the cockpit and infrared camera underneath the fuselage to scan the vicinity for other aircraft as well as assessing weather conditions.
Three off-the-shelf PCs in the cabin are processing all data, with flight control inputs being fed through a Cranfield Aerospace-made unit to the aircraft's autopilot. Data exchange with the ground is provided via an Inmarsat satellite link.
The test team conducted flights over the Irish Sea, during which the system had to respond to intruder aircraft crossing the Jetstream's flight path. The engineers also practised forced landings, where the software automatically selects suitable landing sites. BAE says that the work is essentially a software project to create a "co-pilot in a box".
Dopping-Hepenstal says that the aim of ASTRAEA is not to achieve approval to fly UAVs in general airspace - this he expects to happen toward the end of the decade - but to provide "better understanding and make significant technological developments" on what, he says, is a mutual journey with the regulatory authorities.