Artificial Intelligence in Military Aviation

In mid 2018, the then Minister of State for Defence Subhash Bhamre informed the Lok Sabha that his ministry had initiated the process of preparing Indian defence forces in use of Artificial Intelligence (AI). It shall be used for national security and military strategic purposes and the government was studying a task force report  by a group led by Natarajan Chandrasekharan, Chairman Tata Sons, which had recommended the use of technology in aviation, naval, land, cyber, nuclear and biological warfare. This, he said, had a potential to provide military superiority through both defensive and offensive actions. AI is today a part of research by every think-tank, and subject of seminars and cocktail party conversations. Notwithstanding skeptics, aircrew can find reasons to be enthusiastic about AI.

The development of computer systems to be able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages is termed as AI. Intelligent machine systems that can interpret complex data; perceive the environment and take appropriate actions using learning and problem solving techniques are considered to have AI. Like the human, AI process includes perception, reasoning, knowledge, planning, learning, statistical analysis, computation, and finally manipulates output. It has evolved using expertise in fields like computer sciences, mathematics, psychology, neuroscience, among many others. AI applications already exist in industrial machines, automotive industry, surgery and aviation, among others.

Evolution

          For centuries we have heard fictional stories about thought-capable artificial beings. Even if they did not exist but certainly they were conceptually thought about. The first calculating machine came up in early 1620s which performed based on concepts rather than numbers. Since 19^th century we have seen research on Robotics. In 1940s, Alan Turing’s theory of computation suggested that a machine could simulate any conceivable act of mathematical deduction. Soon digital computers began simulating any process of formal reasoning. Also concurrent discoveries in neurology, information theory, and cybernetics led to consider the possibility of building an electronic brain. The first work that is now is recognized as AI was when McCullough and Pitts’ in 1943 designed “artificial neurons”. Computers then began solving algebraic problems, proving logical theorems and speaking English. Like in most other technologies, US Department of Defense (DoD) allotted huge funds in 1960s for AI. In the early 1980s, AI got a flip with the commercial success of expert systems. Late 1990s AI began to be used for logistics, data mining, and medical diagnosis. Deep Blue became the first computer chess-playing system to beat a reigning world chess champion, Gary Kasparov in May 1997. By mid 2010s, machine learning applications became common across the world. IBM’s Question Answering System defeated the two greatest quiz champions by a significant margin. In 2015 the number of software projects that used AI within Google increased to more than 3,000. Microsoft’s development of a Skype system that can automatically translate from one language to another and Facebook’s system that can describe images to blind people have become possible due AI. In a 2017 survey, one in five companies reported they had “incorporated AI in some offerings or processes”. Around 2016, China  greatly accelerated its government funding; given its large supply of data and its rapidly increasing research output, some observers believe it may be on track to becoming an “AI superpower”. However, some believe that artificial intelligence related reports maybe exaggerated.

AI Show Case Farnborough Air Show 2018

          At the 2018 Farnborough Air show, the latest aviation enabling technologies were the flavor of the season. AI and deep learning which are proliferating beyond the realm of just aerospace were under focus. AI is not only being used for automating the process of testing and analysis, but also all manner of missions on smarter and fully autonomous platforms. AI is already helping evolve autonomous flight control for the electric personal aircraft. The growing unmanned aerial systems (UAS) market will be a great beneficiary, and so would be  Electric Vertical Take-Off and Landing” (EVTOL) systems. Ultimately passengers (troops) and Cargo (logistics support) would be carried autonomously. The air forces are hoping that advances in AI will help access, analyze and process abundance of data from aircraft sensors, weapons and satellites. The data is the ‘renewable oil’ of the 21st century. Streamlined approach to avoid stove-piping information flow which is needed at the speed of war is important. Yet there was a consensus that humans will never be totally out of the loop. Let computers do what computers are good at, and let humans do what humans are good at. This approach could be more workable.

The Network Approach

          A significant amount of collected data never actually gets analyzed as there is low analytical capacity. AI tools will help analysis. A resilient satellite network would be required to provide appropriate effects on a timely scale. Before one gets to AI, there is need for automation by developing trustworthy algorithms. AI will help connect billions of individuals and machines in a nanosecond. When Elon Musk’s Tesla vehicle hits a pothole, it takes into account where the pothole is and its size, and the data is then transferred through network for other electric cars to avoid should they take the same road. Similar exercise needs to be done by connecting all space, air, surface, and sub-surface platforms to pass relevant information about threat adversaries.

Training the AI Systems

          University College London (UCL) team has written what it calls an Intelligent Autopilot System that uses ten separate artificial neural networks (ANN). Each is tasked with learning the best settings for different controls (the throttle, ailerons, elevators and so on) in a variety of different conditions. Hundreds of ANNs would probably be needed to cope with a real aircraft, says Dr Bentley of UCL. To train the autopilot ANNs, data of humans using a flight simulators is farmed. As the plane is flown—taking off, cruise, landing and coping with severe weather and aircraft faults that can strike at any point – the networks teach themselves how each specific element of powered flight relates to all the others.

Evolving New Cockpits 

          Airbus and Boeing forecast future flights without pilots. Unmanned, pilotless technology, will first lead to a single pilot operating commercial airplanes controls, to help airlines reduce costs on crew. Future cockpits would have single wide-screen situational awareness display, using highly accurate 4D (x, y, z, time) full-width, high definition coloured Head Up Displays (HUD) covering entire wind screen including the side windows.  They will overlay flight control, navigation, and low light and IR-enhanced augmented imagery. These displays will be self interpretative and cannot be misread. These tools will allow safe, accurate flight, with reduced  risk of pilot error accidents.

Flight Workload Sharing     

          There is a need for division of roles. The aircraft itself should become a co-operative partner in the aircraft control, flight management and decision-making process. AI will monitor every phase of the flight and ensures absolute compliance and flight safety. In fly-by-wire fighters the computer takes over to prevent exceeding load factor and other flight parameters. The AI compares the stored manufacturer’s exact flight manual performance and handling criteria, and the current flight status. It also takes into account the digitally transferred external situational data and criteria. It then monitors the flight against where it should be. The AI especially monitors all areas where pilot error could affect flight safety. The pilot, of course, still has overall command, but the aircraft shares responsibility, and AI has primary authority for maintaining 100% error-free speeds and altitudes, stall-safe angle-of-attack, and CFIT-safe trajectory. Even in case of pilot incapacitation the aircraft AI could take full control, divert, and land fully autonomously, purely taking instructions from air traffic control.

AI in Aviation Design Support and Flight Safety

          The AI supported Design of Aircraft, or AIDA, is used to help designers create conceptual designs of aircraft. NASA’s Dryden Flight Centre and many other companies have created software that could enable a damaged aircraft to continue flight until a safe landing zone can be reached. The software compensates for all the damaged components by relying on the undamaged components. Integrated vehicle health management system connected to a host of sensors, earlier used in spacecraft, are now used in aircraft. AI machine-learning algorithm would make it possible to cope with unfamiliar situations, such as serious emergencies, sudden turbulence, engine failures, or loss of critical flight data, reducing pilot workload and fatigue. Predictive AI will involve voice alerts. Technologies that help drones avoid terrain, obstacles, traffic, and weather, or self-diagnoses a mechanical problem and  returns to base using AI is already in use. Beyond the alerts, the autopilot could also advise on new flight plans dynamically generated with weather data, fuel consumption rate and other parameters. AI can automatically and periodically run maintenance scenarios and check systems statuses. It can detect failure or inconsistent patterns and trigger repair at the first signs of weakness before something is broken.

AI is thus expected to deliver ultra-safe next-generation airliners and military aircraft – and hopefully finally eliminate human error (HE) in aviation accidents. One approach could be to remove the pilot altogether and design fully autonomous aircraft. Yes, technology is available now. However, are people ready to get onto an aircraft with no pilot? Realistic approach would be combination of both. Keep the pilot in the process of flight, yet take the pilot error out of the process. ‘Smart’ aircraft will use AI and drone technologies to deliver ‘error free’ flight. Current cockpit instrumentation is for pilot interpretation. Most errors are at reading or interpretation stage. Using technology and sensors, the aerial platform knows exactly where it is in all three dimensions all the time. It is this part that needs to be harnessed through AI.

AI in Flight Operations

          Fighter jets are already using some form of AI to provide pilot guidance for best interception profile, weapon guidance and trajectory,  flight departure prevention, and to perform intelligence, reconnaissance, and surveillance (ISR) missions among others. AI has been increasingly used in the F-35 Lightening II aircraft. They are being used extensively in unmanned swarms and mixed manned and unmanned formations. Unmanned helicopters are already landing autonomously on fast moving rocking ships. The new next-generation bomber B-21 Raider could be optionally manned using AI. Russians have been using unmanned MiG-21s as aerial targets for decades through partial autonomous control. The F-16 has an unmanned variant. The current ground control of flight path, sensor payload, and weapons of UAS will soon shift to AI based autonomous flight or be controlled by manned formation aircraft members. This is likely to improve and speed up tactical decision making.

There has been some form of AI for years with autopilots, FADEC, and load-shedding electrical systems all using computer power to make intelligent decisions. These came much before the autonomous cars. Research is on to apply real AI to an autopilot, beyond just programming it to fly certain pre-planned profiles and can adapt to changing aircraft health and environmental conditions. Monitoring and storing hundreds of hours of detailed data from real flights was the first step. A DARPA project called Aircrew Labor In-cockpit Automation System (ALIAS) aims to create a full replacement for a human co-pilot. Initially it will be a mechanical system that manipulates the controls of an airplane like a human does. The next step will be real decision-making tasks.

AI would soon compare current engine signature to a database of millions of hours of engine data. It could also mean ability predicting and warning of impending failure. Predicting weather is another place where AI will be of interest to aviation. AI could decipher the weather radar image and other data to make constructive suggestions. It would analyze fuel consumption rate and winds and suggest flight route. UAS will help developing practical airborne AI. Drones are already doing self-diagnoses of mechanical problem. For regulation and certification of AI, Federal Aviation Authority (FAA) is looking at new capabilities. The pilots would need to be re-skilled to cater for different kind of AI related failures.

Aircraft Simulators & Diagnostics

Airplane simulators are already using AI, interpreting data taken from real flights. AI helps come up with the best scenarios for simulated aerial warfare. AI can also support pilots in the air during combat for instant best attack solutions and maneuver guidance. Interactive Fault Diagnosis and Isolation System, or IFDIS uses a rule based expert system made using collated information from documents and the expert advice. The system allows the regular workers to communicate with the system and avoid mistakes, miscalculations, or having to speak to one of the specialized workers.

Filling the AI Gaps

The system must know when AI is falling short. AI is well equipped for about 80–95% of the task, but remaining needs addressing. AI leaders at Google, Amazon, and the like have figured out that when it comes to mission critical applications, you need a combination of AI and human judgment in order to close the gap. Google Maps was built by using Google’s AI to find the streets and intersections in imagery but then Google’s ‘Team Ground Truth’ (human IQ) had to fill in the gap on tricky one-way streets and construction zones. When it comes to AI systems for complex environment, human judgment is needed to cover the last mile.

The Skeptics – Alternative View

          Some consider AI as a threat to human jobs. Others consider AI a danger to humanity if it progresses unabatedly, and may one day threaten human existence. There are also ethical and morality issues. Mostly discussions have been binary: human intelligence vs. AI. More realistic phrase is “extended intelligence” to signify how AI is used to augment human decision-making rather than replace it.  Some operators are questioning the ‘if-then-else’ approach, that led to the catastrophic Airbus AF 447 crash over Atlantic. The autopilot disengaged in flight because of inconsistent airspeed data input from blocked pitot tubes, exactly as it was programmed to do by its human designers. Dumb programming could be a disaster. But AI is meant to cater to design and input failures by simultaneously taking alternative inputs from other data sources and sensors. Like pilots, the jobs of railroad crew, truck drivers, maritime crew, surgeons, diagnosticians, general management, lawyers, accountants and even government employees could be threatened by AI. Automation in the cockpit has already reduced the pilot to a flight systems manager.

The Future

AI is about ‘man-and-machine’, and not ‘man-vs-machine’. While AI is required to handle myriad tasks with efficient decision making, it needs to interpret and adjust to human emotional state and give an appropriate response for those emotions. ‘Affective computing’ is the study and development of systems and devices that can recognize, interpret, process, and simulate human affects. A motivation for the research is the ability to simulate empathy. Many researchers think that it all will eventually be incorporated by combining all the skills and exceeding human abilities at most or all of them. Anthropomorphic features like artificial consciousness or an artificial brain may be required.

Aviation is one of the most heavily regulated industries in the world due to safety reasons. Strict rules have helped the aviation industry provide the safest way of transport per mile travelled. Aviation incidents are few and far between, and are getting rarer every year – ‘touchwood’. Some degree of automation has indeed helped get aviation to where it currently is. But human control and intervention have always been at the heart of it, from pilots to air traffic controllers. This is about to change. Air freight seems to be the obvious entry point for pilotless planes, just as driverless trucks are about to disrupt the ground transportation industry. When a plane lands, a human alone does not decide how and which ‘gate’ it should go to.

The ever increasing processing power of silicon electronics and the ability to harvest massive amounts of data for computers to process, analyze, and categorize is the single reason for AI to be possible. Amazon, Google, Facebook, IBM, and Microsoft have established a non-profit partnership to formulate best practices of AI technologies, advance the public’s understanding, and to serve as a platform. Fully autonomous AI controlled systems are unlikely to be flying passenger jets just yet. The AI autopilots will probably find its first uses in drones. US DoD is looking into ‘autonomous wingmen’, the Russians and Chinese are looking into fully automated battle fields. Historically, regulation moves slower than technology, because ensuring safety requires lots of tests and certifications. However aviation is ahead of the automotive sector in many areas, thanks to its dynamism.

This Article was earlier published in South Asia Defence & Strategic Review and since then updated 

Picture Source : thedrive.com