Foreign Policy Essay

A Categorical Error: Rethinking 'Drones' as an Analytical Category for Security Policy

By Dave Blair
Sunday, April 24, 2016, 10:00 AM

Editor’s Note: What is a drone? Some do surveillance, others hunt terrorists, and some models likely to enter air forces are more akin to sophisticated fighter aircraft. Dave Blair, badass warrior intellectual, argues that lumping these many different systems together under the label “drone” confuses more than it enlightens. It makes more sense, he contends, to focus on the mission set rather than the engineering behind it.


"All models are wrong, but some are useful."

- Professor George E.P. Box, Empirical Model-Building and Response Surfaces

Poor categorization yields poor understanding, which yields poor policies. A good typology keeps like with like, and distinguishes like from unlike, at least for the purposes of the question you’re asking. In policy, well-built categories help decision-makers appropriately task issues to the most appropriate agency or process, whereas a poorly-constructed typology potentially condemns a complex issue to interagency hell. In the realm of military strategy, the uncertain nature of emerging technologies makes this all the more critical – without placing poorly- understood new platforms alongside well-understood analogs, we risk venturing into the fantastical. And nowhere is this more apparent than in the NatSec debates about ‘drones.’

It is difficult to make coherent policy for the Remotely Piloted Aircraft (RPA) force, because it isn’t a united force in any strategically meaningful sense. Whether you call them Drones, RPAs, or Unmanned Aerial Vehicles (UAVs), we’ve built a category around the remote control system of the RQ-4 Global Hawk, MQ-9 Reaper,* and X-47 aircraft.  This is a clumsy and strange category. The three aircraft have radically different levels of automation. Their missions have little-to-no overlap, and they face entirely different operational and tactical challenges. At best, these planes are cousins; more likely, they are strangers.  The differences between these aircraft well exceed their commonalities, and we would do better to place them in boxes with other craft that do the same sorts of things that they do. 

I recommend that we therefore greatly diminish our reliance on the concept of ‘Drones’ or ‘RPAs,’ and instead lump like with like in terms of mission sets. MQ-9 crews, for example, have much more in common with tactical intelligence, surveillance, and reconnaissance (ISR) aircraft in terms of mission familiarity and skills than they do with more strategic ISR aircraft, such as the RQ-4; RQ-4 crews are fluent in U-2-style surveillance, and vice versa, but know little more than the average pilot about how to fly the MQ-9. Therefore, I propose that we consider RQ-4s as Strategic ISR aircraft, in a (somewhat dysfunctional) family with the U-2, and possibly reconnaissance satellites.  In another example, MQ-9s and AC-130s both spend hours orbiting over a target, staring through advanced sensors and engaging with off-boresight weapons.  Accordingly, we should think of both as ‘Persistent Attack’ aircraft. Finally, X-47s and future Unmanned Combat Attack Vehicles (UCAVs) have more in common with advanced fighters, such as the F-35, than they do with unmanned aircraft like the RQ-4s or MQ-9s. Because of similarities in performance, stealth technology, and mission set, a true UCAV should be considered kin with fighters. These three mission-based categories serve far better than control-system categories for answering questions of strategy and policy for these communities.

Mission-Based vs. Control-System-Based Typologies

The category ‘Drones’ focuses on the control system as if it were the most important aspect of the aircraft. This is essentially tantamount to lumping fighter jets and transport aircraft together because both aircraft use similar electronic control systems or considering transport aircraft like the C-130 and ground attack aircraft like the A-10 as kin because their electro-hydraulic controls. Control systems are one design element amongst many and do not serve as useful analytical categories for most long-term acquisitions or personnel questions. While a control-system-based typology might be useful for technical purposes, it does not help us come to terms with questions of policy or strategy. Accordingly, with the exception of RPAs, the U.S. military does not use control systems, propulsion systems, or crew complement as governing categories for aircraft. Unfortunately, quite a few self-proclaimed ‘drone experts’ unflinchingly accept the ‘drone’ categorization, but lack the technical or operational fluency to cash out the vastly different implications of the vastly dissimilar aircraft and mission sets contained therein. The resulting pronouncements are confusing at best to those of us who actually fly these ‘fly-by-wireless’ planes. 

In fact, outside of RPAs, the Air Force classifies aircraft according to mission set. (The schema for naming aircraft is itself called “Mission Design Series.”) The B-1, B-2 and B-52 bomber aircraft are very different, but they share the bomber mission, and accordingly we placed them into a shared category. C-17s and C-130s transport aircraft have different capabilities, but they share the concept of airlift, and hence, we classify them as cargo aircraft. F-15s, F-16s, and F-22s are likewise differentiable, but they share ‘fighter-ness’ and hence we consider them kin. 

The differences between these aircraft well exceed their commonalities, and we would do better to place them in boxes with other craft that do the same sorts of things that they do.

Changes in tactics and technology can blur category lines – attack aircraft, fighters, and interceptors were distinct classes in the 1950s, but the modern multirole F-16 fighter and the A-10 attack aircraft share trade space. Internal politics, too, can use the taxonomy to skew policy outcomes: power struggles between the Air Force’s “Bomber Mafia” and “Fighter Mafia” resulted in the nearly identical F-111 and FB-111. This typology has worked well for the Air Force, as bomber acquisitions strategy or fighter manning forecasts are both meaningful concepts. The mission-based taxonomy keeps “like” together and separates “unlike” well enough to make meaningful decisions. Therefore, I propose that we reconsider which aircraft are close kin and which are distant cousins with respect to mission sets, rather than control systems. 

By placing remotely piloted platforms under a control-system-based exception to this taxonomy, we’ve made it very difficult to think about all of these platforms. The MQ-9 dominates the discussion with ubiquitous lethal strikes, but this is not the reality of the RQ-4 world, which primarily consists of strategic ISR. The ‘drone’ chip-out from the mission-based typology confusingly conflates concepts between these largely disjointed worlds; by decoupling these discussions, we can engage each with far more analytical precision. This conflation also confuses the MQ-9 discussion by overly focusing on the control system and thereby neglecting commonalities with manned aircraft of similar mission sets. While ‘Remotely Piloted Aircraft’ is a more accurate and precise term than ‘drone,’ it still makes the same categorical mistake. I propose we rectify this mistake by leaving the control system category to the electrical and computer engineers, and instead, looking toward categories based on mission sets.

How We Got Here

Perhaps more than anyone else, the current category owes its coherence to the superb work of Dr. Tom Ehrhard, who described the history of UAVs in a story arc from the remote-controlled B-24s of the Second World War to the modern Predator. However, we are victims of his well-earned success – the discussion, especially inside the Air Force, has heretofore taken his lineage and typology of our current RPAs as writ. This provenance places the high-performance, jet-powered, Vietnam-era RPAs, such as the BQM-34 Firebee, as progenitors to the well-known MQ-1 Predator, a claim that has generally gone unchallenged despite the differences in performance and mission between the two craft.

While this lineage was helpful for a time, it has outlived its usefulness. What was once an incubator is now a straitjacket, and the rapidly-maturing MQ-1/MQ-9 and RQ-4 communities are increasingly in a position to chart a course well beyond these initial concepts, which will likely chart courses largely independent of each other.

Now that we have overcome the novelty of ‘fly-by-wireless’ aircraft, it is time to re-focus on what it is they do. And they do many different things, some of which they do very similarly to their manned counterparts, and sometimes, very differently from each other. Three extant mission categories capture the spectrum of capabilities that we currently, clumsily discuss as ‘drones’: Persistent Attack, Strategic Reconnaissance, and Air Dominance. 

By placing remotely piloted platforms under a control-system-based exception to this taxonomy, we’ve made it very difficult to think about all of these platforms.

Persistent Attack: Gunships, Tactical ISR, Predators and Reapers

The MQ-1 Predator is a far more banal flying machine than all the talk of ‘drone revolution’ would let on; the craft is little more than a motorglider with a datalink, a camera and a missile. What is actually revolutionary is the ability to stare at a target while parked overhead for hours at a time and engage that target at a time and place of your choosing. And that revolution predates the Predator by decades. 

Aviation pioneer Abe Karem gave the Predator wings, but airpower pioneer Colonel Ron Terry gave it claws. Colonel Terry developed persistent attack airpower in order to suppress the Ho Chi Minh trail during the conflict in Vietnam. His basic theory held that an aircraft that could loiter – and look and shoot from that loiter – could exert vertical dominance over the ground so long as one could hold the air. His theories took form in the AC-47 and the AC-130 Gunships, which fused three characteristics: the ability to park over a target, observe it with a sensor package, and engage from that perch with off-boresight weapons. This combination proved devastating in conflicts from Desert Storm to Iraqi Freedom. 

Colonel Terry’s vision was expansive: he imagined swarming remote aircraft partnering with manned gunships in a persistent cloud over that battlespace. He built an early version of that cloud over the Ho Chi Minh trail, with early versions of that vision realized in the partnership between early AC-130s, OV-10 spotter planes, and QU-22 unmanned intelligence gathering aircraft. This vision bears a striking resemblance to the partnership of persistent ISR/attack aircraft overhead modern battlefields.

These aircraft share bloodlines. The camera ball on the MQ-1 and the camera on the U-28 were both developed out of an effort to develop a new targeting camera for the AC-130 Gunship during the 1990s. Many of the people who founded the modern tactical ISR community, both manned and remote, came from the Gunship circa 2005.  Enlisted sensor operators hailing from the AC-130 Gunship community transformed the culture of the early Predator community from an ISR collection community to a tactical raid support and strike platform. All three of these communities share far more DNA with each other than any of them share with the RQ-4 or a future UCAV. Moreover, the aspects of the MQ-1/MQ-9 that have reshaped the spectrum of policy options – endurance, precision strike, and sensor fidelity – are about persistent attack, not about robotics. Therefore, the concept of persistent attack should guide us toward better manpower and acquisitions choices and ultimately, a better future for the community. 

[T]he concept of persistent attack should guide us toward better manpower and acquisitions choices and ultimately, a better future for the community. 

Strategic ISR: Global Hawk & U-2

The dominance of the MQ-1/MQ-9 in the ‘Drones’ discussion has, unfortunately, drowned out the voice of the RQ-4 Global Hawk community. Their story is distinct and better understood if we consider them as siblings with the reconnaissance U-2. The RQ-4 and the U-2 have broadly similar sensors and mission payloads, performance profiles, and mission sets. They both work closely with intelligence ground stations that direct and analyze their payloads of exotic sensors. They face many of the same challenges and meet those challenges in many of the same ways. They should be considered as kin. The friction between the two communities is then a sibling rivalry, with each focusing on how their particular concept of operations outdoes the other. Future acquisitions choices will eventually put this rivalry to rest, but as interesting as that story will be, it is orthogonal to the issues faced by MQ-1 Predators and MQ-9 Reapers. 

Air Dominance: F-22, F-35, X-47

Lastly, the X-47 fills a very different role than either the RQ-4 or the MQ-1/MQ-9. A UCAV is best understood as a remotely-piloted or autonomous strike fighter aircraft. With high performance, long-range sensors, stealth technology and so on, these craft are built with the hallmarks of advanced fifth-generation jet fighters and should be considered as such. While this technology is still maturing, and while we await a synthesis between the current crop of manned fighters and the possibilities of these new technologies, the UCAV seems on track to become kin, partner, or rival to manned fighters. This drama will certainly be fascinating for aviation aficionados, but it once again has little to do with the future of persistent attack or strategic ISR.

Next Steps: Quad-Copters and Small, Swarming Unmanned Aircraft Systems (UAS)

The contemporary discussion generally distinguishes between Small UAS (sUAS) and larger, military-grade UAS. These are two different creatures: in a mission-set category realignment, we should retain the instinct to separate out the sUAS and Quadcopters, store-bought mini-UAVs and their kin, as something new and different from their military counterparts. The Department of Defense explicitly differentiates these classes, as does the very helpful typology from the Center for a New American Security. Given size and cost, Small UAS has no manned counterpart and can have no manned counterpart. As these technologies link with better swarming algorithms and more efficient datalinks, we should expect to see even more disruptive and fascinating applications come from these aircraft. 

I recommend referring to any automated craft too small to have a manned equivalent as a ‘drone.” By this standard, the word ‘drone’ actually suits these aircraft well – they fly through simple algorithms, they mostly take care of themselves, they can accomplish simple tasks on their own. Similarly, the emerging class of indefinite-duration, high-altitude, solar-powered flying wings seems to be carving a novel niche as ‘surrogate satellites.’ Since there is no manned equivalent of a craft with one-month or longer endurance (except perhaps the International Space Station), these would be something new as well and might rightly be called ‘drones’ or something else entirely. Either way, that discussion is largely uncoupled from the fates of the other three categories of aircraft discussed here. 

Conclusion: Electrical Engineering is not Foreign Policy

The technologies that supposedly create the ‘drone revolution’ animate all of the next generation of tactical aircraft, both manned and remote. The F-22 blows the MQ-1/MQ-9 away in terms of processor speed. The F-35 is no slouch in the realm of datalinks and networking.  And frankly, the automation that goes into both of those aircraft puts the MQ-1 Predator and MQ-9 Reaper to shame. We make an analytical error when we overly focus on computers and datalinks in remote aircraft, but overlook them in manned craft. Therefore, the term ‘RPA’ has a place in technical aeronautical and electrical engineering discussions, but much less so in foreign policy circles. Just as someone who cannot explain feedback loops has no business opining about ‘fly-by-wire’ aircraft, a working understanding of latency should be a prerequisite for those who wish to publish serious works about remote aviation. Fortunately, mission-based categories require much less of a technical preamble, and they draw upon operational concepts familiar to students of foreign policy. 

We clarify our thinking about policy and strategy when we focus on the capabilities and mission sets of aircraft. For instance, we could easily speak about shared future sensor and weapon acquisition equities between the MQ-9 Reaper and the AC-130 Gunship. For almost all questions that spark policy debates, “persistent attack” or “strategic ISR” bins these questions correctly and focuses on the most important feature: operational employment. And this is much of what we, as a community, have been saying all along: what you do should matter more than from where you do it. 


*For the purposes of this paper, I consider the iconic MQ-1B Predator as subsumed by the MQ-9A Reaper, because they share deep similarities and were at one point considered two variants of the same craft.  In fact, some of the sub-menus in the MQ-9 refer to the aircraft as a ‘Predator B’ rather than a ‘Reaper.’

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of the U.S. Air Force, the Department of Defense, or the U.S. Government.