Maverick’s next wingman might be a drone.
Make that four or five collaborative combat aircraft, or CCA, working together to execute a set of “plays” with minimal direction from a human pilot, who “quarterbacks” the mission from a formation positioned near those drones.
That’s the vision of the U.S. Air Force for its sixth-generation fighter, and one of the service’s seven operational imperatives for the future force, specifically “Defining the Next Generation Air Dominance (NGAD) System-of-Systems.”
In March 2022, during the Air Force Association’s Warfare Symposium, Secretary of the Air Force Frank Kendall said during his keynote address that “NGAD must be more than just the next crewed fighter jet. It’s a program that will include a crewed platform teamed with much less expensive autonomous, uncrewed combat aircraft, employing a distributed, tailorable mix of sensors, weapons and other mission equipment operating as a team or formation.”
Unlike today’s uncrewed aircraft, which are remotely piloted from the ground, tomorrow’s uncrewed aircraft will be robotic teammates that make decisions for themselves. They’ll autonomously aviate, communicate and navigate, and then carry out one or more reference missions in collaboration with each other and at the direction of a human mission commander.
Combat pilots, however, and commanders who send them into action must trust that these CCAs — “loyal wingmen,” as they’re often called — will perform as expected, following the rules of engagement, adhering to mission principles, effectively contributing to the mission and, most importantly, behaving safely. Before the service will ever consider fully adopting and integrating this new capability into combat operations, they must perform predictably and appropriately.
For combat pilots and commanders to gain trust in the technology, they need to be included in the design loop from the very beginning. Whenever a new autonomy requirement or design reference mission is needed, engineering teams must bring current operators and pilots into the process, getting their input in early technical discussions.
The system engineering process should start with questions for operators, not prebaked answers: What do you want these things to go do? What is the objective they need to accomplish? How long should it take? What data do you want to collect? How do you want to interface with the robots? How should their information be displayed?
By including operators early and throughout the entire process, it aids in establishing performance requirements. Plus, by including operators from the start, there’s a stronger willingness to trust the autonomy.
Today’s remotely piloted platforms are operationally challenged in highly contested environments, where they can face layers of adversarial defenses or communications jamming. In these environments, collaborative autonomy is necessary to ensure the autonomous aircraft can work with other tactical aircraft to provide the mass and maneuver required to win the fight.
As has been stated by senior Defense Department officials many times, autonomous teaming aircraft could include a family of capabilities, ranging from less expensive and attritable drones to more expensive, exquisite designs, all intended to complement current and future crewed combat aircraft, as well as each other.
Defense companies are working to find the right points on the cost-capability curve for each instantiation, but everyone seems to agree that assembling this potent mix of various capabilities is how we win.
It’s also expected that this approach will facilitate faster design upgrades. I believe it’s these new collaborative autonomous aircraft that will provide the best opportunity to create and field the long-vaunted Digital Century Series.
These new CCAs could carry out a variety of missions like reconnaissance, electronic warfare and the suppression of enemy air defenses. And since the crewed and uncrewed aircraft are collaborating and communicating with each other, the team can identify threats and targets quicker than a lone tactical platform.
CCAs also act as a force multiplier by extending the sensor range and combat radius of sorties, allowing a smaller number of crewed aircraft to cover larger areas of the battlespace. Furthermore, by mimicking the performance, capabilities and radar cross-sections of crewed aircraft, CCAs could overwhelm and confuse adversaries by acting as decoys to draw fire away from crewed teammates.
To conduct these types of demanding missions, CCAs will need some level of artificial intelligence. As the technology matures, machine learning could further aid in rapidly and automatically improving algorithms for decision-making and courses-of-action development. However, a fully autonomous uncrewed aircraft powered by AI and machine learning will require rigorous testing and certification — something several years away.
Instead, in the near future, I envision CCAs will use a more deterministic set of decision-making capabilities for specific missions. This type of “weak AI” has been performing remarkably well during development and flight testing. It still provides significant collaborative capability amid rapidly changing, complex environments, but it is more contained and better defined, capable of executing common, planned tactics just like today’s human wingman. Importantly, it can also be tested, certified and, ultimately, trusted faster.
In the near future, I don’t believe it’s reasonable to expect our nation’s leaders to unleash hundreds of swarming drones to learn and collaborate using a giant neural network in a highly contested environment. It’s much more likely you’ll see mixed teams of humans and machines, complementing each other’s capabilities, where the machines are provided just enough learning and decision-making autonomy to accomplish specific aspects of the operational mission.
Tom Cruise might be getting older, but the real-life Mavericks out there won’t be hanging up their flight suits anytime soon. While we can hope to augment our “Top Guns” with new, autonomous, collaborative aircraft and capabilities, we still need their expertise in the air. And they need to be able to trust their wingman, whether it’s a human or robot.
Ted “Hefty” Conklin is the technical director at Raytheon Technologies’ intelligence and space division.