Warplanes: Autonomous F-16 Flies and Fights


May 24, 2024: The U.S. Air Force recently demonstrated an autonomous (no pilot) F-16 using flight software enhanced with AI (Artificial Intelligence) that demonstrated it was able to fly combat maneuvers better than most human pilots. Moreover, a pilotless fighter would be smaller and capable of maneuvers that a human pilot could not handle. Human pilots cannot handle certain violent maneuvers because the G (gravity) forces cause the pilot to black out. Unmanned aircraft only have to worry about violent maneuvers that might cause the airframe to be damaged, perhaps enough to cause a crash.

For decades commercial and military aircraft manufacturers have been seeking ways to add more automation. For example, in 2019 the U.S. Air Force demonstrated a new system that could turn any manned aircraft into an unmanned aircraft, controlled by a system of electronics and mechanical devices that literally replaces the human pilot in the cockpit. Best of all, ROBOPilot can be quickly removed so a human pilot can fly the aircraft again. Installation takes a little longer, but not by much because the device is designed to be available quickly should the need arise.

What is novel about ROBOPilot is not that it can replace a human pilot, but that the ROBOPilot equipment is flexible enough to be quickly installed in any manned aircraft. ROBOPilot software has to have an electronic profile of each aircraft it can be installed in and has to be tested to ensure the ROBOPilot mechanical components can use all the controls of that aircraft. That said, cockpits on aircraft are pretty standard. The most complex and eccentric cockpits are for fighter aircraft and the air force has decades of experience building custom versions of what became ROBOPilot for several generations of jet fighters.

The first test aircraft for ROBOPilot was a fifty year old Cessna single-engine commercial aircraft. These aircraft don’t have a lot of the electronic controls modern aircraft use. While electronic controls are easier for ROBOPilot, the most difficult aspect of developing ROBOPilot is creating a robotic arm and hand to handle the many tasks a human pilot has to handle. The 1960s Cessna was mostly mechanical controls and analog displays for instruments. Moreover the Cessna cockpit is cramped compared to more modern aircraft. The Cessna, in effect, was a more challenging test for ROBOPilot than a larger more modern aircraft. But the Cessna was also cheaper to use as a test aircraft for the first flight, from a remote airfield in a thinly populated area, in case all the safety devices failed and there was a crash. There were no problems and the entire two-hour flight was recorded on video from inside the aircraft.

The ROBOPilot project then moved on to see how difficult it would be to adapt the basic ROBOPilot system to other aircraft. The goal is to have a ROBOPilot kit that aircraft maintainers can, with a minimum of training, install in any of the aircraft ROBOPilot has a profile for. This includes test flights and regular profile upgrades to match changes made to each aircraft’s flight controls.

ROBOPilot was developed, in part, because of a larger Loyal Wingman program that uses drone/UAV (Unmanned Aerial Vehicle) fighters, or other types of aircraft, for particularly dangerous missions. That idea was the outgrowth of more than half a century of efforts to automate cockpit functions. By the 1970s it was possible to create robotic pilots for fighter aircraft. Back then that had some very critical practical applications. The main purpose of remotely controlled fighter aircraft is so they could be used as realistic aerial targets. The equipment used to convert retired fighters into robotic ones have become more capable and reliable since the 1970s and part of that effort turned into the successful development of automatic landing systems for commercial aircraft, which are now a common item that takes a lot of stress out of landing an airliner at night or in bad weather. This software has even been adapted to land jet fighters or naval UAVs on aircraft carriers at night.

Another approach to this automation appeared in the form of PIBOT, which puts a robot in the pilot’s seat. The robot is designed to handle the flight controls like a human but do it under AI control. Tests so far have been successful and the test and development program is scheduled to be completed by 2026 with versions available for commercial and military aircraft. PIBOT can also be adapted to operate trucks and other vehicles. That means military truck convoys could use PIBOT to replace all but one or two of the human drivers.

ROBOPilot and PIBOT can also be used for the Loyal Wingman program. In early 2017 the air force demonstrated that F-16s equipped to operate as UAVs could successfully operate in formation with manned F-16s. This was an important goal for Loyal Wingman, a program for eventually integrating combat UAVs with piloted warplanes. The F-16 UAV needed software that would allow it to fly in formation, execute attack missions on its own and avoid interference from jamming. That software worked although the initial flight tests of Loyal Wingman simply confirmed that the F-16 UAVs could safely fly in formation with piloted F-16s and effectively receive and respond to commands from the flight leader or other piloted F-16s.

Work is continuing to develop software that will enable the F-16 UAV to carry out complex attack runs on its own. This involves avoiding ground fire from missiles and using its own EW (electronic warfare) equipment to deal with jamming. All this live software testing would eventually be used in combat UAVs like the ones the navy has been testing and the air force is now developing. The U.S. Army has already been testing similar software control of UAVs by suitably equipped with secure digital communications gear used by attack helicopters.

Loyal Wingman came about after four years of effort to develop a UAV version of the F-16. This UAV version was based on the QF-16, the remotely controlled target version of the F-16. Back in 2013 the air force got its first QF-16 flying and began converting over a hundred retired F-16s to QF-16s. At the time it was noted that with a little extra work the QF-16 could be turned into a combat UAV for dangerous missions like SEAD (suppression of enemy defenses) or attacking ground targets guarded by heavy air defenses. The air force was already planning to use combat UAVs for this but those are not available yet. It was noted that QF type aircraft use GPS to help with navigation and to ensure that QFs flying in formation don't collide with one another. The QF-16 also carries sensors to detect near misses by missiles. Out of that came modified software and some additional hardware to enable the recent flight tests.

There are several other reasons for adapting the F-16 to be a combat UAV, in addition to designing combat UAVs from scratch. For one thing, the UAV version of any combat aircraft is superior in some ways to one with a pilot in it. This is mainly because pilots black out when the aircraft makes turns too sharply at high speed. QF-16s used regularly for Loyal Wingman training also make it possible to monitor via additional sensors in the QF-16, how extreme for piloted aircraft maneuvers stress the fighters. Tests of manned aircraft have always been part of the development and testing process for new and modified aircraft. But those tests were limited by the cost of losing a lot of aircraft. By the 1960s fighter aircraft design had reached the point where the aircraft could perform maneuvers, like high speed turns, that humans could not handle. The air force discovered how effective this capability was during the 1970s when they rigged some jet fighters to fly without a pilot and had them go up against manned aircraft. The remote-controlled fighters were able to outmaneuver the same aircraft carrying pilots. It was a scary demonstration of what was to come.

The QF-16 has already demonstrated its ability to carry out acrobatic maneuvers under remote control. This would be useful in getting into heavily defended air space. Adding more sensors and flight control software could produce a formidable combat UAV. Even when all the QF-16 conversions are completed there will be several hundred retired F-16s suitable for conversion to combat UAVs.

The air force ordered the first QF-16s in 2010 and the initial flight test took place in September 2013. The QF-16s are converted from existing F-16s that have been retired from service. The air force ordered 125 QF-16s and deliveries began in 2015. Each QF-16 conversion costs about $1.2 million and consists of installing hardware and software that enables remote, no pilot in the cockpit, flight control. The process of equipping the F-16 with all the necessary sensors like cameras and remote feeds of the aircraft radar, and remote capabilities took longer than expected, even though there was a lot of experience doing this to older aircraft like F-4s, F-100s, F-102s, and F-106s. The original plan was to introduce the first QF-16s in 2011. The QF-16s can still carry a pilot who can fly the aircraft or simply observe how the remote control process works.

The QF-16 replaces the older QF-4 drone aircraft. Nearly 250 F-4 Phantom fighters were modified to fly by remote control. The mods cost about $1.4 million per aircraft. The QF-4 first appeared when the U.S. Air Force retired its F-4 fighters in the 1980s. The air force has run out of retired, but still flyable F-4s to convert and QF-4 conversions ceased in 2013.

Training operations destroy up to 25 remotely controlled QF class fighters a year. As the supply of decommissioned F-4s was exhausted, the QF-16s arrived just in time. Before the QF-4, the air force had converted 218 F-100s for use from 1983-92, 136 F-102s from 1974-85, and 210 F-106s from 1990-98 to act as full-scale target aircraft. There are smaller UAVs that are also used as targets. The full-scale models are needed to fully test the capabilities of new, and existing, missiles. Nothing like using real missiles against real targets to build pilot confidence and be sure the missiles will work in combat.

There are so many retired F-16s available that there are plenty that could be used as combat UAVs. The F-16 manufacturer is not doing UAV conversion research, but rather another aircraft company, Boeing, which sees a potential market for such aircraft. These UCAVs (Unmanned Combat Aerial Vehicles) already exist as the MQ-1 Predator and MQ-9 Reaper. But an MQ-16 would be cheaper. This would be an inexpensive way to see what a more ambitious and larger UCAV could do.



Help Keep Us From Drying Up

We need your help! Our subscription base has slowly been dwindling.

Each month we count on your contributions. You can support us in the following ways:

  1. Make sure you spread the word about us. Two ways to do that are to like us on Facebook and follow us on Twitter.
  2. Subscribe to our daily newsletter. We’ll send the news to your email box, and you don’t have to come to the site unless you want to read columns or see photos.
  3. You can contribute to the health of StrategyPage.
Subscribe   Contribute   Close