Simulations and Flight Tests of an Automatic-Ground Collision Avoidance System (Auto-GCAS) for Performance Limited Aircraft

摘要

Controlled flight into terrain (CFIT) is a leading cause of aircraft total losses and fatalities among Department of Defense (DoD), commercial, and general aviation communities. Despite the F-16 Automatic-Ground Collision Avoidance System (Auto-GCAS) fielding in 2014, there has been minimal transition in Auto-GCAS algorithm design from the single-escape high-performance maneuver aircraft to a multiple-escape limited-performance maneuver aircraft-defined by MIL-STD-1797 as large, heavy, and low to medium maneuverability. Using a stitched Learjet-25D model (LJ-25D), an Auto-GCAS algorithm was developed to predict multiple-escape maneuvers through the trajectory prediction algorithm (TPA), compare these trajectories to a safety buffer added to digital terrain elevation data (DTED), and trigger at the last open aggressive escape maneuver upon predicted terrain collision. Simulations against a modified DTED terrain map highlighted Auto-GCAS design refinements in forward look-ahead times and trigger handback criteria. Compared to the baseline LJ-25D Auto-GCAS algorithm, extending the forward look-ahead time past 45 seconds did not decrease collision prevention performance, but changing the trigger handback from a first-open to forward-open requirement reduced collisions against vertical cliff walls by 17.5%. Additionally, time-varying forward look-ahead times increased prediction awareness at lower air speeds without added computational cost, allowing the Auto-GCAS algorithm to successfully escape a box canyon where previous methods failed. 10 flight tests at the U.S. Air Force Test Pilot School (USAF TPS) demonstrated the ability of the LJ-25D Auto-GCAS algorithm to control the Calspan Learjet 25D Variable Stability System (VSS) safely away from ground collisions. The required algorithm refresh rate was 12.5 Hz to ensure zero ground collisions, corresponding to a required minimum forward look-ahead time of 18 seconds. The LJ-25D Auto-GCAS algorithm was able to achieve and maintain the target aircraft load factor and flight path angle with momentary overshoots. The algorithm trigger and handback showed no tendency for nuisance activations, proving this as a viable operational Auto-GCAS system for performance limited aircraft.