A lightweight automatic ground collision avoidance system (Auto GCAS) which depends on a terrain database of the entire world housed in a smart phone is being flight tested by NASA.
The flights follow initial development of the Auto GCAS for U.S. Air Force F-16, F-22 and F-35 fighters, and are aimed at extending use of the safety system to a much broader range of applications from unmanned air vehicles to general aviation aircraft. The system uses precise navigation, performance and digital terrain data to constantly monitor the aircraft’s position relative to known obstacles. Auto-GCAS is designed to intervene as a last resort, automatically recovering an aircraft when it senses that a ground collision is imminent and the pilot is not taking action.
ACAT testbeds - F-16D and Super Flyin' King model aircraft. (NASA)
Unlike the original Automatic Collision Avoidance Technology (ACAT) Auto GCAS project which was flight tested on an F-16D in 2010, this new test phase is using a 9 ft 8 inch wing span radio-controlled hobby Super Flyin’ King model aircraft. Tests of the Droid (Dryden Remotely Operated Integrated Drone), are being undertaken by NASA Dryden Flight Research Center with support from the Office of the Undersecretary of Defense and the Air Force Flight Test Center.
“We’ve talked about the modular architecture which means you can put it on other aircraft, but we’d only flown it on an F-16. So we asked why not try it out on a very different aircraft,” says Auto-GCAS project manager Mark Skoog. Aside from demonstrating the transition and portability of Auto-GCAS to another platform, the project is also designed to develop and explore performance enhancements such as new escape maneuvers beyond the vertical fly-up used by the F-16 system.
New lateral escape maneuvers (marked in blue) are being tested in the latest flight test phase. (NASA)
Primary objectives include integrating the safety device without increasing the weight of the UAV, demonstrating a system that functions autonomously during a lost link with a ground station and showing the GCAS algorithm can be run using a smart phone – in this case a Motorola Droid X. The phone uses the open Android operating system architecture which can be modified to run the Dryden-developed Auto-GCAS software.
“We’re flying with the entire globe modeled on the phone. We can take it anywhere on the planet and it will have collision avoidance capability,” says Skoog. “This shows you can break the paradigm on accessibility to other platforms. You don’t have to go through a full-scale Defense Department development cycle, and it opens up many opportunities in the future,” he adds.
A 'Droid' being readied for testing at NASA Dryden Flight Research Center (Guy Norris)
Pushed by the unique needs of the F-22’s avionics architecture to develop data compression techniques for the fighter’s terrain database, the team were able to springboard off this and squeeze a digital version of the entire surface of the planet into an Android. “We were able to go to 2,000:1 compression ratio,” says Skoog who adds that by converting the earth’s surface into a tile-based elevation database more than 300 gigabytes were compressed to 145 megabytes. “This is the first time we’re flying real time de-compression, long before the F-22. They’re very interested in our flight test results.”
For the Phase 1 test the phone was integrated into the ground station rather than the UAV and connected to the aircraft’s autopilot through an interface unit via a command and control link. The system’s capabilities were assessed during an initial set of 10 flights conducted between late September and October in rugged terrain near Edwards AFB, Cal. For each flight, the UAV was flown against a cliff face or up a blind valley by a pilot in the ground station with a safety pilot monitoring from close by.
For the upcoming second phase the phone be integrated directly on board with the autopilot, a relatively unsophisticated Piccolo 2 unit. Although limited by the slow 70kt speed of the aircraft, tests will continue to focus on system performance, trajectory prediction and possibly upgraded scanning logic which will look further out. “We also plan to fly with zero terrain clearance buffer before the end of the program,” says Skoog
Droid pilot control station (Guy Norris)
Aimed at the terrain by the pilot, the vehicle executed automated lateral escape maneuvers using algorithms modified to reflect turn rates and trajectories based on the performance of an MQ-9 Predator. Following each escape the safety pilot resumed control.