In the cockpit of a fifth-generation fighter, somewhere high over the ocean, an Air Force pilot gets a heads-up: hostile fighters are airborne and closing. The call didn’t come over the radio. Instead, the pilot’s cockpit display shows detections made by a flight of unmanned aircraft operating far to the west. A pair of low-observable, highly capable adversary fighters is outbound at high speed from their home coastline.
The pilot and the other human crews on station know this because their unmanned vanguard is watching from multiple perspectives, using multiple sensors. One UAS used its infrared search-and-track sensor to spot the hostile aircraft as they climbed out from the coastline; a second one, cued by the first, slewed its sensors onto the track and confirmed the presence of the contacts from a different angle.
Artificially intelligent and autonomous systems aboard these and other aircraft know from the location of the launch, the speed and nature of the contact, characteristic recognition and other features, that these fighters are of a particular type belonging to a hostile air force. They also know to escalate this information to their human commanders immediately, along with an interrogative: What should they do?
The pilot, the UAS, and the other aircraft operating in the area, can’t be sure at any given moment which of their satellite-based systems might work. Those are regularly degraded by hostile activity. But the UAS and the human-occupied aircraft nonetheless can stay in touch with each other, including via mesh networks in which each aircraft is a node. So, every human pilot in the area gets an alert immediately after the autonomous aircraft systems conclude the conditions are right.
With that information in hand and time to act, fighter pilots on station can refer to their orders, the standing rules of engagement, or higher authority in deciding what to do. They might retrograde and try to open up distance from the bogeys. They might energize their own radars or use other sensors. Or they might authorize the UAS in this rapidly approaching future scenario, some of which are armed, to fire.
Unmanned systems already have transformed warfare, intelligence, and other national security applications. They’re doing so again for information dominance and air combat, making possible near-future scenarios like this one. To be clear, the new age of UAS applications to air combat won’t look very much like the past or even current operations. Today, when an MQ-9A Reaper takes off for a mission, it’s under the control of a human crew in a ground control station responsible for the safe flight and mission. The aircraft is advanced and automated, but it still receives focused attention by many human specialists, all at once, for long periods of time.
Artificial intelligence and autonomy mean that tomorrow’s UAS will do much of their work on their own. Operators essentially will push a button that makes them take off and fly their missions. They’ll sweep their own sensors over the ocean surface or through a box of airspace and make their own discoveries.
In the old days, it might have taken a human intelligence specialist to look at the sensor feed from a UAS and say, ‘hey, that ship claims on the Automatic Identification System that it’s a civilian trawler, but I can see that actually it’s a hostile surveillance vessel.’ The subject matter expertise of a human intelligence officer was required to add value to the pictures and other surveillance captured by the UAS. Starting now, and growing into the future, a UAS will make this kind of assessment for itself. It will contrast what’s being shown on AIS with what it perceives and what it knows about the configurations of the vessels in its operating area and conclude: This is an anomaly. I’d better let someone know.
GA-ASI’s advanced software and integration systems are supporting these kinds of detection and analysis capabilities today, providing users with the highest quality intelligence, surveillance, reconnaissance ever possible from unmanned platforms. That will only improve. And in the new era, UAS will share what they observe and what they conclude and be poised to act when circumstances require. If newly independent aircraft operations are the first big innovation required for the new era of air dominance, the second builds from it: resilient and dynamic communications.
U.S. and allied warfighters must prepare for combat against sophisticated adversaries capable of degrading legacy networks. Today, if a hostile actor jammed the satellite communications of a Reaper, that wouldn’t hurt or crash the aircraft, but it could stop its remote human operators and other consumers from observing the area until it was clear of the disruption. Tomorrow, more resilient, low earth orbit satellite communications will make these links harder to disrupt. New point-to-point laser signaling systems won’t be susceptible to distant radio frequency jamming. And mesh networking within an allied force means that UAS and human-piloted aircraft can transport and broadcast their own communications with them.
Picture a flight of friendly aircraft operating in a column many miles long. In the lead are new-generation UAS; trailing are human-piloted fighters and high-value patrol aircraft, such as the P-8 Poseidon or the E-3 Sentry. As the lead ships in the column cross into an area in which satellite communications are degraded, the aircraft begin to rely on signals beamed in all directions from and among their peer aircraft. A trailing aircraft outside the jamming envelope can preserve a satellite link, if necessary, and share out data within the network.
So new UAS won’t need nonstop positive human attention to operate, because they’ll fly themselves and operate their own mission systems. Call it supervised autonomy. They don’t need always-on satellite or radio networks. They can escalate and burst-transmit only when necessary. And when they do need to interact with the battle force, new technology and new tactical communications strategies mean they can. A lead-element UAS accompanying a flight of fighter jets may see several air and surface targets over the course of their patrol, but they know to escalate onto the network and alert only when they detect specific military units under pre-specified conditions.
This is what enables the third innovation that will revolutionize UAS operations: the preservation of human control in the most critical situations. Although the chain of events may be different, the ultimate authority for the release of weapons or other important mission events still will be up to humans. Even though a supporting UAS might carry air-to-air weapons, only the human fighter pilot – or another human commander – would authorize their use. These concepts of operation for newly autonomous and intelligent UAS are about more than building the technology, networks and software required. They’re also about integrating what’s now becoming possible into the legacy practices of human warriors. Think of it like training a cavalry horse or a working dog – then training humans to work with them to the greatest effect.
Warfighters must be able to interact with these advanced systems in a way that doesn’t distract them or overload their already considerable responsibilities as combat aircrews. That requires a simple-yet-powerful interface. And above all, the humans in the battle force must trust the machines. Trust comes from the net total of experience – humans learned to trust elevator buttons to take them to their floors of choice, or parking payment kiosks to raise the swing arm blocking exit from a garage. Military pilots and air crews can learn to trust new AI and automated systems via manned-unmanned teaming.