Is the Era of Manned Warplanes Over?

The 'pilotless' F-16 fighter aircraft with an empty cockpit has been flown by the USA for test purposes in recent years. In reality the pilot; He flew the F-16 from the ground, from the control station. At this point, the idea that we can now fly existing F-16s 'without a pilot' has not been pronounced by anyone. Because F-16s are a whole with their pilot, even the training level and experience of the pilot means a value above the capabilities of the aircraft. A trained fighter pilot is an essential component of the F-16.

Flight tests like the pilotless F-16 have shown that similar warplanes can be flown 'without pilot' if equipped with appropriate flight control systems and mission computers. However, it has also been realized to a large extent that the combat missions expected from an F-16 cannot be carried out with ground control. If artificial intelligence (Artificial Intelligence - AI) finds a place for itself in the cockpit of the aircraft as a "virtual pilot", it is likely that warplanes will make autonomous (non-pilot) flights in the near future. However, in the near term, it is highly unlikely that autonomous UAVs, whose development is on the agenda, will be able to successfully perform combat missions in the three-dimensional, complex air theater. The basic need for success is to be able to model the fighter pilot, not the F-16, but the mind, training level and experience of the pilot.

What should be understood by the word UAV?

Unmanned Aerial Vehicles (UAV) is a type of flying vehicle that does not physically contain humans. Today, UAVs are produced in many different shapes, sizes, configurations and characters.

Fully manned or fully unmanned aerial vehicles have several inherent advantages and limitations:

A manned aircraft can be safely used on the battlefield, in controversial environments where command and control is limited and political constraints exist, under conditions where precision is required, by giving decision responsibility to the pilot in the loop within the framework of the rules of engagement.

There is no air crew to limit the range and durability of an unmanned aircraft. There is no loss of life or fear of being caught. Comparatively, they are less costly aircraft than manned platforms. It is easier to make up for their losses.

The most important and indispensable component of UAVs is the ground control station (YKI), in which UAV pilots sit. In other words, this is the remote cockpit of UAVs. It is essential to have two-way communication between YKI-UAV. Uninterrupted and bi-directional control 'inputs' via radio-link, on a specially allocated, appropriate frequency channel/band (with/without crypto), within line of sight (technically up to 200 kilometers) to the antenna, via radio-link. Flights are carried out by transferring the outputs and necessary data. Similar to the use of mobile phones, by using satellite instead of rado-link, YKİ-UAV communication can be carried out uninterruptedly in the coverage area of ​​the satellite.

For example, AKSUNGUR, which flew at the TEKNOFEST Aviation, Space and Technology Festival held in our country between September 21-26, 2021, was flown by the pilots at the ground control station in Ankara, and the 'spectator greeting flight' was successfully carried out at low altitude over the festival area.

The "weight" of UAVs is determined by a classification made among themselves, but such classifications are not binding. According to the aviation literature, the UAV classification accepted as a generic classification is as follows:

Tactical UAVs:

Tactical UAVs are low-speed, relatively small, low-weight UAVs built to fly up to 18 000 feet above ground altitude. GNAT 750 etc. belongs to this class.

Operative (Medium Altitude Long Endurance - MALE) UAVs:

UAVs in the middle class, built to fly between 18 – 30 000 feet, are called Operative UAVs. Such UAVs can carry more payloads and cruise relatively faster than tactical UAVs. Bayraktar TB2, Karayel, ANKA, AKSUNGUR, AKINCI, MQ-1 Predator and MQ-9 Reaper are among the UAVs we can add to this class.

Strategic (High Altitude Long Endurance - HALE+) UAVs:

They are UAVs with long range, which are built to fly at 30 000 feet and above, and can stay in the air for a long time. Generally, they are UAVs that serve the purpose of intelligence, reconnaissance and surveillance rather than armed missions. RQ-4 Global Hawk, RQ-3 DarkStar etc. UAVs are included in this class.

Special cameras, targeting pods (including FLIR pods) can be attached to UAVs, and weapons and ammunition can be integrated. All such extra systems or weapons are known as payloads. In this case, if we want to name the UAVs carrying weapons and ammunition, we can say Armed UAV or SİHA. However, there is no such concept as SİHA in aviation literature. Just as airplanes are not differentiated between armed and unarmed, such a nomenclature for UAVs does not exist in aviation jargon. But somehow, Turkish defense and aviation In the heating industry, the nomenclature of SİHA has been adopted in the relevant media organs. It is widely used.

UAV Production and Use in Turkey

In the Turkish Aviation Industry, significant developments have been made in UAV production in recent years. Turkish Aerospace Industries (TUSAŞ) with its ANKA and AKSUNGURs, Baykar Makine with Bayraktar TB-2 and AKINCI, and Vestel Defense with Karayel are prominent Turkish companies in this field. Rotary wing UAV manufacturer DASAL's KARGO-150 offers a wide range of solutions for civil and military use. Numerous domestic companies that produce other small or mini UAVs have also made significant progress in their own fields.

Within the scope of Turkey's internal security and cross-border operations needs; The use of UAVs in intelligence, reconnaissance and surveillance activities has increasingly continued since the 1990s. In this field, the US-made GNAT and the Israeli-made Heron have been used by the Turkish Armed Forces until the mid-2010s, when Turkey's own UAVs came to the field. To the extent that national solutions can be developed for the logistics problems that arise within the scope of the maintenance and maintenance of these aircraft, it has been possible to carry out successful missions, especially with Herons, in internal security and cross-border operations. Afterwards, Turkish-made UAVs, with cameras and ammunition payloads produced by Turkish companies, can be used for SAR, electronic warfare, etc. when necessary. It started to work by being equipped with systems and received the flag from the Herons. UAVs, which are our own production, have accomplished successful missions beyond our borders. He continues to sign.

Use of Artificial Intelligence (AI) in Classic UAVs

Almost all of the UAVs currently produced and used in the world use artificial intelligence to some extent. Autopilot, automatic landing and take-off system, mission planning, recognizing the image or shape of the target, etc. areas can be given as an example of the use of artificial intelligence in UAVs.

However, to the best of our knowledge, an autonomous UAV that can only fly and serve with artificial intelligence, and that can perform a self-flying task without a pilot in the operation of the system, has not been produced yet. In the current situation, there must be a human in the loop in UAV flights. In addition, firing the weapon carried on the UAV or throwing/leaving the ammunition may not be left to the decision of artificial intelligence under current conditions, without the pilot or the utility load operator being involved. Even if technology provides this opportunity, current rules of engagement, duties, powers and responsibilities have not yet enabled the use of UAVs in this type of war modeling.

Flying aircraft with advanced programming methods based on advanced software using previously known data is a programming activity. It would not be correct to consider it as an autonomous flight. If all the necessary data and intelligence flow are provided to the engineers responsible for programming, in a timely manner, the next thing will be nothing but the operation of the 'programming' logic.

If what is meant by artificial intelligence is that the aircraft systems react according to the previously known modeled world, the movement surfaces are activated, the missile is fired, etc. However, in this case, it seems possible in the near future to fly conventional UAVs 'autonomously' without humans in the loop and to have them do a certain amount of work.

Use of Artificial Intelligence (AI) in Warplanes

Warplanes don't fly, they fight. Flying warplanes like classical UAVs is against the purpose of the airplane and the nature of the war environment. By fighter aircraft we mean here, multi-role, air defense (hunt), bombardment, air defense pressure, tactical reconnaissance, close air support, support of naval operations, etc. It is an aircraft capable of performing missions.

A 'predictable' flight profile such as taking off from an aerodrome, flying its planned route from a certain altitude in level flight, following the cleared descent profile and landing on a planned or reserve aerodrome with the passenger plane algorithm, would not be suitable for the philosophy of use and mission requirements of fighter planes.

Again, in a predictable world, having warplanes with a known profile, for example an air show, cannot be considered as autonomous flight. An F-16-like fighter jet can be rendered unmanned and made to perform, for example, soloturk-like flight demonstrations without a pilot in the loop. Because this is an engineering model of a known world. The algorithm is clear. According to the loaded sequential program, the F-16 can perform the demonstration flight. Taking this type of flight does not mean that fighter planes no longer need a pilot. Because, even the theoretical dimension of the use of artificial intelligence, which can correspond to the pilot's mind that should be in the loop, has not been fully resolved yet.

There are decades of work by aviation industry companies to explore ways to use artificial intelligence to improve the capabilities of warplanes. As with UAVs, differences in usage patterns 

However, developing early versions of artificial intelligence are already used as an auxiliary system in warplanes. For example, digital flight controllers and mission computers are early applications of artificial intelligence that ease the workload of the 'fighting' pilot who flies in the cockpit every day.

In the current situation, the expectation from artificial intelligence systems is as follows: By activating an algorithm that will not tire the pilot with unnecessary information, it will enable the pilot to fly his aircraft effectively and to perform the task expected from him, in a 'complex air combat environment', according to the requirements of network-centered air operations. The artificial intelligence needs to pass the outputs or data of the integrated systems through a filter and transfer as much as necessary to the fighter pilot (audio, visual, command alert, etc.). Rather than confuse the pilot with the data, in a way, artificial intelligence applications and software as a whole should act as if the plane were a virtual co-pilot.

The normal pilot receives support from the systems for maneuvering and suitable firing positions while maneuvering according to engagement priorities in the air communication environment. Considering the position, altitude and distances of the target aircraft, the artificial intelligence should give the necessary directions to the pilot in a timely manner so that the pilot can carry the aircraft to the optimum point in the three-dimensional plane. By putting the helmet-mounted sight system-like overhead screen displays and gun shooting parameters in the pilot's eye, artificial intelligence systems can accurately guide the pilot.

In addition, a pilot who is the leader of the task package has to carry out additional command and coordination duties in addition to flying his own plane and performing the task assigned to him. The artificial intelligence systems on the plane should support the leader pilot's ability to be an 'conductor' in the airspace. For example, it enables the aircraft in the arm to use the air battlefield synchronized and engage the enemy aircraft with suitable weapons and parameters, to provide direction guidance for safe route and avoidance maneuver against surface-to-air missiles, to activate chaff/flare-like protective systems, etc. Artificial intelligence-based systems should show the signs that will warn the pilot.

All these issues are already present in today's modern warplanes. In this context, progress continues. In today's advanced aircraft, especially in 5th generation warplanes, without artificial intelligence support, in some cases without the introduction of autonomous sensors and systems integrated into the aircraft, it has become almost impossible to use all the systems of the aircraft effectively with only the pilots and to fulfill the combat missions expected from that aircraft.

However, with the support of all 'early' artificial intelligence applications in question, it is not possible to fly warplanes, to have some missions, and to claim that the requirements of the definition of autonomous flight are fulfilled.

Autonomous War UAVs

When we say combat UAV, what we understand is, for example, the pilotless version of the F-35 aircraft. It should be understood that the F-35 aircraft can perform its flight duties with at least as much efficiency as the F-35 with real pilots, with the autonomous combat UAV without humans and pilots in the loop.

In a flight, that flight mission can be considered autonomous if the artificial intelligence can perform the task safely, effectively and economically, without humans in the loop, at all phases of the flight, from starting the engine to muting and then to the completion of the flight mission critique.

For this, there is a need to move from the early artificial intelligence applications based on the 'modeling of the world' perspective to the 'modeling of the mind or brain' phase.

In other words, the moment we reach a point of confidence that we can entrust our country to the engagement decisions of a 'virtual pilot' who can take the place of the war pilot we have given flight training, experienced and taught all the rules of engagement to protect our country by teaching them to the battlefield to protect our country. We can start using it.

Within the framework of Autonomous Warfare UAV development studies, we consider that it would be beneficial to look at civil and commercial applications for modeling the mind or brain, which are ahead of the developments in military aircraft.

In recent years, we have witnessed that countless companies have come to the stage of producing 'driverless' or autonomous cars by spending billions of dollars, and successful test drives have been carried out in this context. Promising and very successful projects for the commercial sector excite humanity. However, there are no 'driverless' or autonomous taxis, buses, etc. yet. we are at a point far from being seen on our streets.

This is because current autonomous vehicle projects have not yet met the "security requirement" to ensure that we can all get from one point to another safely. Fully autonomous vehicles called Level 4 (no need for a driver) We know that it has not yet successfully passed the autonomous driving tests. Still, the rollout of autonomous cars with no humans or drivers in the loop is likely to happen soon.

Transferring the autonomous usage experiences of land vehicles moving in a single plane to aircraft flying in a three-dimensional environment will contribute to paving the way for important advances in the field of military and civil aviation. There will be many lessons for autonomous UAV manufacturers to learn from the process of developing autonomous land vehicles.

As a matter of fact, studies on the production of war UAVs, which are candidates to replace warplanes, are carried out by aviation industry companies. For example, in July 2021, the USA pitted a piloted fighter jet with an autonomous combat UAV for 'air combat'. A method similar to Kasparov's playing chess was applied in 1997 with the IBM-manufactured special computer (Deep Blue) capable of trying 200 million positions per second. The flight tests of autonomous war UAVs together with piloted warplanes have been a good start for opening new pages in the field of war UAVs. It is known that the USA aims to reach the necessary design data for the production of truly pilotless aircraft flying with artificial intelligence at the end of this project.

With similar studies, we continue to progress rapidly towards the autonomous flights of artificial intelligence systems in an integrated and integrated structure. The ability of combat UAVs to perform autonomous, pilotless flight missions is a matter of technology readiness and time. Besides, in today's warplanes, there are things that have an answer with pilot experience and training level. The most challenging missions for Autonomous Warfare UAVs are counter air operations missions performed at enemy depths. At the same time, the autonomous decision-making process must be carried out rapidly in a three-dimensional environment in missions such as air defense, protection, hunting scans that can attack enemy aircraft in the air.

It is a priority issue to find an answer to how situational awareness and on-the-spot decision-making needs can be met by autonomous warfare UAVs in such complex, friend-enemy intertwined environments, which are unknown due to the nature of warfare. This is an area that UAVs are not yet designed to overcome. In particular, according to the available data, according to the war environment analyzed and interpreted by the human brain, decisions must be made on a second-by-second basis by the same human brain. Artificial intelligence to be used in autonomous warfare UAVs will need 'brain cloning or replication', which will ensure that it makes decisions as fast and accurately as the brain of a trained fighter pilot.

Autonomous bomber, transport, tanker, strategic reconnaissance, surveillance, intelligence, etc. UAVs

In the near future, it may be possible to autonomously perform some military flight missions (many of which resemble civilian air transport) within the framework of a pre-programmable algorithm based on 'world modeling'. AKINCI or AKSUNGUR can be made autonomous in this sense.

As a result, such tasks are performed in a two-dimensional rather than three-dimensional (plane) environment, to put it simply. Such flight missions are not performed in a three-dimensional, highly variable and threatening environment. Mission types included in this title are carried out in low-threat environments, far behind the main battle line. Thus, in a world where everything is predictable and pre-programmed, these flight missions can be performed for military or civilian purposes. Autonomous UAVs with the capabilities to perform these tasks can be produced. There may be no need for humans or pilots to be involved in the loop.

The point to be emphasized here is that even in cases where people are in the loop, side damage can be caused or friendly elements can be damaged in UAV attacks. Whereas, the damage done to friendly elements as a result of such bombings in manned warplanes is almost negligible. Autonomous UAV designers will also need to produce solutions for this issue.

The crux here is the difference between modeling the world and modeling the mind. The current level of technology cannot take us to the 'autonomous' structure in warplanes, where the mind needs to be modeled. UAVs in statements such as "we will build a pilotless warplane" or "the era of manned warplanes is over" by some executives at the head of aircraft manufacturer companies or people like Elon Musk, who is known for his futuristic works; On the other hand, relatively 'simple' UAVs operating in the type of this title, yes, it is possible to produce pilotless, fully autonomous UAVs suitable for such missions in the very near future.

In this respect, there are countless advanced solutions in the field of civil aviation, from air transportation to delivery of cargo loads to addresses.

We are going through a period where he is entering a crisis. There is no reason why there should be no parallel military solutions.

The Limiting Factor for Combat UAVs: Electronic Jamming/Assault Environment

Electronic warfare, especially electronic jamming and the use of UAVs on the territory of countries with developed offensive capabilities is quite risky. As a result, the control inputs of the UAV are currently sent by the pilot via radio-link and/or satellite, and the accuracy of the pilot commands is confirmed by the returned data. All this process is operated in real time or negligibly delayed (near real time). In the event that it is flown in an environment that will 'confuse' the commands to be given to the UAVs, it is only a matter of time before the UAV, which is subject to interference, becomes inoperable. If an algorithm has not been defined for the aircraft to return to the take-off aerodrome or the closest friendly abode using its own navigation system, the UAV may crash. Therefore, UAVs can be used more effectively and safely in low threat environments, against countries with weak or no electronic warfare capabilities.

Even if the UAV is stealth (invisible), in any case, the radio frequency will be able to make even such low-visibility UAVs 'visible'. For this reason, it has already become a necessity to go for autonomous solutions in war UAVs.

In addition, UAVs; It is not yet ready for dogfight or close combat. Although the first air-to-air missile from UAVs was launched in 2002, 20 years have passed, sending UAVs into enemy territory without gaining air superiority means that you will be ready for countless UAV losses. Every resource is valuable in war, even if there is no human in it. Every aircraft, even unmanned, is a valuable power element. Moreover, the developing concepts are based on the unmanned autonomous flying of aircraft as expensive as the F-35. An operation plan that ignores the loss of the unmanned F-35 is probably unthinkable.

Conclusion

It is considered that a minimum of 15-20 years is probably needed for an autonomous combat UAV. As a matter of fact, the development of both manned and unmanned versions of 6th generation warplanes has been put on the agenda by some countries or consortia. In any case, if progress can be made in the field of modeling the mind, it may be years before the pilotless planes we see in futuristic movies will protect our skies.

Abbreviations in the article:

AI: Artificial Intelligence

DASAL: Aselsan-Altınay Partnership, UAV production company

GNAT: General Atomic (US made tactical UAV, in use since 1989.)

HALE: High Altitude Long Endurance

UAV: Unmanned Aerial Vehicle

MALE: Medium Altitude Long Endurance

TAI: Turkish Aerospace Industries Inc.

YKİ: Ground Control Station (station used by pilots and payload operators to fly UAVs)

Some references that we used in this article:

Rincon P. (2020), “Combat drone to compete against piloted plane”, Science, BBC News, 5 June, <https://www.bbc.com/news/science-environment-52933958>, p.e.t.27.9.2021.

Chopra A. (2013). “Manned vs Unmanned”, Dassault Aviation, August Issue, <https://www.sps-aviation.com/story/?id=1278>, p.e.t.27.9.2021.

Lee H. (2021). “F-35 Pilot: Forget Drones, The Skies Still Belong To Fighter Pilots”, June 14, <https://www.sandboxx.us/blog/f-35-pilot-forget-drones-the-skies-still- belong-to-fighter-pilots/>, set 27.9.2021.