The F-16 GCAS is a modified block 25 F-16D to perform flight tests of an automatic ground collision-avoidance system. This system has demonstrated that the use of advanced computing technology can significantly reduce the number of accidents attributed to controlled flight into terrain (CFIT). The U.S. Air Force has lost 4-5 aircraft per year to CFIT since the early 1990s, and the Swedish air force has a CFIT rate about twice that. However, the GCAS system now being developed could reverse those trends.
A USAF, Lockheed Martin, NASA and Swedish air force Advanced Fighter Technology Integration (AFTI) team completed flight tests of an automatic ground collision-avoidance system or Auto-GCAS on an upgraded USAF Block 25 F-16D in the fall of 1998. In 29 flights, the team conducted more than 350 test maneuvers - such as diving at the ground and the side of mountains - to fulfill two key objectives of the program:
- Demonstrate that an Auto-GCAS could significantly reduce critical fighter-aircraft mishaps resulting from pilot spatial disorientation, loss of situational awareness, G-induced loss of consciousness (G-LOC) and gear-up landings.
- Identify any areas where an Auto-GCAS might impede a pilot's performance of standard tactical missions.
During a dramatic 1-hr. demonstration flight for this it became clear that these objectives essentially have been satisfied. The system is not mature enough to install in production fighters yet, but it's about 95% ready.
There is no funded program in place now, but the Air Force will probably field some type of Auto-GCAS on the F-16, F-22 and Joint Strike Fighter. Air Combat Command is developing a formal requirement, and there appears to be enough top-level interest in curtailing perennial CFIT accidents that the AFTI team's research won't be relegated to a dusty shelf, a senior USAF officer said.
The Swedish air force, which routinely flies missions down to 100 ft. above mountainous terrain, could be the first to install a production Auto-GCAS on their JAS 39 Gripens. Sweden's Forsvarets Materielverk (FMV) and Saab participated in Auto-GCAS development and flight tests, and cofunded the program with the U.S. Air Force Research Laboratory under a collaborative agreement. Four Swedish air force pilots have flown Auto-GCAS demonstration profiles at Edwards AFB, and their assessment of the system's viability was very positive.
I found the system to be a sophisticated and complex--but quite robust and effective-- last-ditch method to save an "unaware" or unconscious pilot's life. LMTAS and Saab engineers blended GPS/inertial navigation inputs, a digital terrain database, a radar altimeter, and the AFTI F-16s autopilot with an Aircraft Response Model (ARM) to create a full-envelope, automatic ground collision avoidance system. The block 50
F-16 Terrain-Referenced Navigation System provides a "position" input to help a new algorithm decide what nearby terrain could present a hazard to the aircraft. Based on the fighter's maneuvering attitude at any given moment, a specific area of the terrain database is "scanned," and elevation information is compressed into a 2D model.
The ARM is a sophisticated simulation of the F-16, running at a real-time rate. "It's a fairly complicated algorithm that tracks fuel-burn, takes information from the stores management system [about weapons weight and drag], and even accounts for system processing delays," said Mark A. Skoog, USAF's AFTI F-16 test director. "Using the aircraft's current state, the ARM computes a full six-degree-of-freedom simulation during a roll to wings-level. At wings-level, [ARM switches] to a 2D-type recovery--a second-order modeling of the jet's pitch response. It calculates how much [kinetic] energy it can trade for altitude until the jet reaches a desired zoom-climb speed, then holds that speed."
Ultimately, the computer determines how much time is available before the aircraft will break through a pilot-selected minimum descent altitude (MDA), then triggers an autopilot- commanded protective maneuver. Typically, two chevrons (><) appear in the head-up display 5 sec. prior to an automatic fly-up, warning the pilot that Auto-GCAS is about to take over. If he takes no action, the chevrons move toward each other until their points meet, a flashing "break-X" symbol appears in the HUD and an aural warning annunciates, "Fly-up! Fly-up!"
At that instant, the Auto-GCAS commands some of the most aggressive, eye-watering maneuvers this ex-USAF flight test engineer and civil pilot has ever experienced.
If inverted (bank angle greater than 90 deg.) and somewhat nose-down, a negative 1g push throws the pilot "up" into his shoulder straps and lap belt to get the aircraft's nose headed skyward. Immediately, a 180-deg./sec. roll is commanded, bringing the aircraft to wings-level, right-side-up. Somewhere after passing the 90-deg.-bank point, a 5g pull-up is initiated at an approximately 4g/sec. rate. The system commands a maximum angle-of-attack recovery, if flight conditions will not sustain a 5g pull-up. When the flight path is pointed above threatening terrain, the Auto-GCAS disengages and announces, "You got it!"
If aircraft speed is insufficient to climb adequately during a pull-up, the system also will announce, "Power! Power!", urging the pilot to push the throttle forward and gain airspeed.
Skoog's system briefing prior to my flight emphasized the speed of these autopilot-performed lifesaving maneuvers, and the reason for their aggressiveness. "The roll has a very rapid onset. It'll impress a spot on the side of your head. Pilots will get their heads banged against the canopy. But for a sole-purpose Auto-GCAS, there's no reason to be delicate about the roll [rate]. Let's get to wings-level as quickly as we can," he said.
"The roll onset rate is faster than any pilot can command," Prosser added. "There's no way you can move your hand that fast to command a sharp onset. It's so fast that you get the impression the roll rate is much higher than it is."
Because pilots are adamant about having final authority over their aircraft, the AFTI test team initially gave the pilot an ability to always override the Auto-GCAS. Extensive testing, plus discussions with F-22 test pilots, changed that attitude.
"During all-terrain testing, we found that even the slightest override of the GCAS autopilot in the wrong direction would blast you through the [MDA] floor," Skoog said. "Trying to do elevated-g fly-ups, we saw hundreds of feet of additional altitude loss due to a 0.3-sec. override. We came out of the program knowing that we'd have less protection by giving the pilot total autopilot override [authority]. So, we lock-out the pilot in roll and yaw. He can add pitch up to the angle-of-attack limits," and can always deactivate the Auto-GCAS by hitting a "paddle" switch at the base of the control stick, or pushing with 19 lb. of force.
For our flight, Prosser set the Auto-GCAS in "Active" or "Low-Level" modes, with "Standby" appearing on the HUD only when the landing gear handle was moved to the gear-down position. He activated the system shortly after our takeoff from Edwards AFB in F-16D serial no. #81176, a Block 25 aircraft that had its avionics updated to a Block 40/50 configuration. As the first digital F-16 testbed, this airframe was operated under several structural limitations that restricted maneuvering somewhat.
Prosser initially demonstrated GCAS response rates by setting up a shallow dive, then pulling the gun-trigger on the F-16s sidestick controller to initiate a pilot-activated fly-up (PAFU). I repeated the PAFU check from a near-inverted, 30-deg. nose-down attitude, verifying that depressing the back-seat gun-trigger also would command a fly-up maneuver. This PAFU technique was used throughout the Auto-GCAS flight-test program to command an immediate pull-up for any safety-related reason.
A team of engineers in a mobile ground station monitored airspeed, dive angle, bank angle, Auto-GCAS status, time-to-fly-up, and distance-from-terrain data telemetered from the jet. Special computer displays in that control room enabled an immediate "abort" call if engineers saw any aircraft parameter exceed preset limits.
The test pilot also could initiate a PAFU if he "exceeded his comfort level," Prosser said. Extensive testing has shown that pilots typically initiated a pull-up about 1.5 sec. before they would hit the ground.
"We had to have pretty tight tolerances for these test runs," Prosser said. For one that simulated a blacked-out pilot in a steep dive, "that meant we only had a 10-deg. dive, 20-deg. bank and 50-kt. airspeed [tolerance]. When you're pointing straight down, that's pretty hard to nail."
During Prosser's final demonstration of an aggressive PAFU-commanded recovery from an initial 120-deg. bank, 17-deg. nose-down attitude, our spare digital data transfer cartridge (stowed in the front cockpit) smashed through the map case cover due to the sharp negative 1g push-up. The cartridge--slightly larger and much heavier than a VHS videotape cassette--hit the canopy, flew aft toward me, then, during the positive-g pull-up, slammed onto a panel between our cockpits. It was an unexpected and graphic illustration of the high accelerations and forces commanded by the AFTI F-16 autopilot.
After moderate checks of the system at shallow dive angles and an aborted run or two, Prosser simulated several fatal mishaps. The first replicated a pilot flying on night-vision goggles (NVG) and losing situational awareness. With Auto-GCAS minimum descent altitude set at 500-ft. AGL (a medium-risk test condition), Prosser rolled into a partially inverted 5g turn, then back to a 90-deg. bank before relaxing his grip on the stick. The mishap pilot had lost the night horizon and, thinking he was approximately wings-level, let the nose fall. He was unknowingly diving toward the ground. Similar NVG-related accidents have killed F-16 and A-10 pilots.
While the flat Rosamond Dry Lake raced upward at us, filling my out-the-canopy field-of-view, I glanced at my back-seat HUD repeater and saw two large chevrons moving toward the center of the display. Their arrow-points touched, and we immediately snap-rolled to wings-level and pulled sharply to about 10 deg. nose-up. When the "You got it!" annunciation sounded, we were climbing at about 317 kt. and 2,940 ft., roughly 600+ ft. above the lakebed--an artificially high altitude established for safety reasons.
"[Auto-GCAS] just saved your life," Prosser said.
USAF F-16D block 25, #83176
, used for the combined US-Swedish GCAS trials, held in 1998 (USAF photo)
The next mishap simulation depicted a pilot dropping a low-drag bomb in a 20-deg. dive, then pulling up at 5g, rolling into a 135-deg. left bank and looking over his shoulder to watch the bomb's impact. Instead of climbing to downwind, however, the unwary pilot allowed the aircraft nose to drop about 20 deg. below the horizon.
We initiated the test run from a base altitude of 8,300 ft., simulated a bomb release at 375 kt. CAS and 5,500 ft., and started a 5g pull-up. Prosser rolled 135-deg. left and let the nose drop. However, our dive angle rapidly increased to 28 deg, prompting an abort call from the ground-based control room and a pilot-initiated fly-up via the PAFU switch about 1 sec. early. Local air-traffic conflicts precluded a repeat run.
Prosser reset the Auto-GCAS MDA to 50 ft., selected "ground speed" for display on the HUD and descended into a preplanned low-level tactical course. To avoid unnecessary distractions, he eliminated the chevrons from the HUD, as well. We flew about 200 ft. above the ground at 520-560 kt., popping over high-tension power lines, hills and small ridges. Slipping through cuts in the desert mountains, rolling inverted to pull down the backside of ridges, and carving around the sides of rocky hills, Prosser demonstrated that a pilot could fly a normal, low-level tactical mission without experiencing a single nuisance fly-up.
"So far, [Auto-GCAS] has not impeded our mission at all," he noted.
When we were inverted, the system lost radar altitude information, which double-checks aircraft location by comparing the ground's contours along the flight path with those stored in the digital terrain database. However, the sophisticated Auto-GCAS algorithms compensated appropriately, relying on inertial system-derived altitude information when necessary.
The test pilot reset our MDA to 200 ft. and flew at a constant altitude across a peak to demonstrate the slight "speedbump" effect of grazing that minimum-descent altitude. The Auto-GCAS "fly-up" was very brief and docile--just enough to let a pilot know he was close to the protective altitude band separating his aircraft and the terrain. Prosser had restored the chevrons to their normal 5-sec. prefly-up appearance schedule, which warned us of an impending "bump" as the peak approached.
Because Fremont Peak was just beyond, we experienced a second brief "Fly-up! Fly-up!" call when Fremont came into the system's scan pattern a second or so after getting the first "speedbump" fly-up. The second one was almost imperceptible, though.
Before the next two runs--directly at the steep slopes of what the test team dubbed "GCAS Mountain"--Prosser explained the euphemism, "pilot comfort level."
"We'll have the flight path marker in the dirt, so [these] will engage my [personal] comfort level," he said. A test pilot's judgment about how long to wait before squeezing the PAFU trigger was a significant factor in Auto-GCAS development, and personal differences had to be accounted for in the data processing.
Our first run was a wings-level, 465-kt. approach to GCAS Mountain with a 700-ft. MDA set. The HUD's flight path marker was aimed about 2/3 up the peak, and the desert rocks, dirt and scrub brush raced rapidly at our windscreen. About the time I would liked to have suggested "Pull NOW!", the system took over and the 5g-plus pull-up drove me down in the seat. The g-onset was so rapid, then sustained, that the oxygen mask threatened to slide over my nose, and the g-suit tried to compress my lower body to half its normal size, it seemed. The aircraft nose reared up to a 28-deg.-high deck angle before the "You got it!" call sounded. The radar altimeter showed we were 1,040 ft. above the mountain, and our speed had dropped to about 400 kt.
"That one was pretty close to my comfort level," Prosser quipped. "I wouldn't have gone, maybe, 2 more seconds," which was about 2 sec. beyond the Aviation Week guy's comfort level.
The next run was one of the most-impressive--a 30-deg. dive directly at the side of GCAS Mountain. Prosser set the Auto-GCAS MDA to 2,000 ft. (considered a low-risk test MDA), climbed to 7,500 ft., rolled inverted and pulled the nose down to -25 deg. He snapped the fighter back to wings-level and pushed the nose further down to hold a 30-deg. dive angle, aiming at the rock-strewn mountain. The fly-up annunciation came at 2,750 ft. AGL and around 400 kt., triggering a long, 5.3g pull-up.
Even test pilots met their match when flying this test point during the development effort. None of the program's pilots would go lower than a 100-ft. MDA.
Next, Prosser set a 50-ft. MDA and descended to 100-150-ft. AGL for a low-level run over fairly smooth terrain at about 500 KTAS. Again, there were no nuisance fly-ups as we raced across the desert floor along Cords Road. Prosser pointed our flight path marker at the base of Desert Butte, a lone mountain north of Edwards AFB. The jagged mountain filled our windscreen--and extended well above it--before we heard "Fly-up! Fly-up!" Just as the high-g pull-up relaxed, another fly-up was triggered, possibly because Prosser had hit the PAFU trigger almost coincidentally with the Auto-GCAS activation.
Prosser concluded the mishap demonstrations with a simulated G-LOC situation, emulating a pilot who pulled enough gs to pass out. Its pilot unconscious, the fighter eventually would fall nose-down and accelerate.
With a conservative 6,000-ft. test MDA set, our first attempt was aborted when airspeed exceeded test limits. On the second try, Prosser rolled into a 120-deg.-bank, 5g turn at 20,250 ft., then relaxed, letting the nose drop. Even with speedbrakes extended and power at idle, we hit 530 KTAS and were 55 deg. nose-down in a roughly 110-deg. bank when the Auto-GCAS triggered a fly-up around 12,200-ft. MSL. The system then had to overcome our approximately 900-ft./sec. descent rate. A data check indicated our recovery bottomed out 280 ft. above the MDA.
"We were milliseconds away from hitting the 'ground,'" when the Auto-GCAS took over, rescuing the unconscious pilot, Prosser said.
On final approach for landing at Edwards, the Auto-GCAS switched to "Standby" as soon as the gear handle was lowered. If we had made a touch-and-go, the system would have automatically reactivated at gear-up. For safety reasons, it also is automatically deactivated when the air refueling door is opened, or alternate flaps are selected.
About the only key improvement that needs to be made to the Auto-GCAS tested here is installation of faster microprocessors and, possibly, an alternate navigation "solution," Skoog said. Even then, no processor-related problems were noted during flight evaluations until the F-16 approached 800 KTAS, well beyond where most missions would be flown in proximity to hazardous terrain.
The Air Force and its Swedish partners are laying out follow-on efforts, but those familiar with the recent test program agree this AFTI F-16 team has completed the necessary development, and done it safely. Without question, they have lived up to the program's Swedish motto: "Du Kan Inte Flyga Lagre," which translates to "You Can't Fly Any Lower."