In This Issue
Centennial of Aviation
March 1, 2004 Volume 34 Issue 1

The Role of Size in the Future of Aeronautics

Monday, March 1, 2004

Author: Alan H. Epstein

Once Aircraft no longer required human operators, they could be made much smaller.

Since the first controlled powered flight by the Wright brothers 100 years ago, aircraft size and range have been important measures of progress in aviation. Large payloads have always required large aircraft (although the definition of large has evolved over time). In the early days of aviation, the fuel capacity required for long-range flight also dictated large size; thus, size and range were coupled. The first aircraft sold by the Wright brothers in 1909 to the U.S. government, the Wright B Flyer, had a 40-foot wing span, a takeoff weight of 1,400 pounds, a range of 90 miles, and required a pilot and an observer. In terms of payload and range, the B Flyer was a state-of-the-art airplane.

Ten years later, the first aircraft to cross the Atlantic Ocean, the Navy-Curtiss NC-4, had a wingspan of 126 feet, a crew of six, and a takeoff weight of 27,000 pounds. This airplane was designed by NAE member Jerome Hunsaker. Because the NC-4 had a range of only 1,400 miles flying at 75 miles per hour, it had to be refueled along the way. By the time of Lindbergh’s solo nonstop transatlantic flight in 1927, aeronautical technology had progressed to the point that an aircraft with a wingspan of 46 feet and weighing only 5,100 pounds could fly nonstop 3,600 miles from New York to Paris. In one sense, the Lindbergh airplane had an optimal design - the flight control, navigation, and payload were embodied in one person. Twelve years passed before the next transatlantic milestone, the start of scheduled commercial service in 1939 by the 84,000 pound Boeing Model 314 Clipper with a 152-foot span and a nominal range of 5,200 miles at 184 miles per hour. By this time, large size was dictated as much by commercial considerations (basically seat mile cost) as by technical considerations. Commercial considerations continued to dictate the size of aircraft throughout the jet era.

Eighty-four years after the Wright brothers’ flight, technology had advanced sufficiently that the Aerosonde, a model airplane with a 10-foot span and weighing about 30 pounds, could cross the Atlantic. Lindbergh’s heroic, lone pilot was replaced by inexpensive avionics, a microprocessor, and a satellite navigation receiver. Like the NC-4, the Aerosonde cruised at about 75 miles per hour. In 2003, the even smaller TAM-5, a model airplane weighing just 11 pounds, made a similar transatlantic flight. In contrast to this smallest transoceanic flyer, we now have commercial aircraft weighing more than a million pounds with transpacific range. Thus, by the end of the first century of flight, the connection between aircraft range and size that had been so important in the early years of aviation had been broken.

In military aircraft, range and size have been important historically, but military planners now focus on achieving effects, such as destroying a building. Figure 2 presents a historical perspective of the number of sorties (one flight by one aircraft) flown to ensure a hit on a 60-by-100 foot building. In World War II, B-17s flew more than 3,000 sorties. A B-17 required an aircrew of 10 and at least that many in the ground crew, so more than 60,000 people were involved, a true army of the air. By the time of the Vietnam War, only 44 sorties were required to hit a 60-by-100 foot building. In the latest Iraq war, only one was required. So the 750,000 tons that took flight in 1944 were reduced by a factor of 30,000 to 25 tons.

Because the number of aircraft on a mission cannot be reduced below one, further reductions in aerial mass for military missions must be in aircraft size. The bombers I just discussed all had gross takeoff weights of about 50,000 pounds. How large must an aircraft be to get the job done? Consider, for example, the task of destroying a well defended target, such as a bridge. In 1972, this was typically done with F-4 Phantoms, an expensive proposition given the inaccuracy of unguided bombs and aircraft loss rates. One contemporary analysis put the price at $12 to $15 million, not including the cost of necessary "extras," such as combat air patrols, tankers, and defense suppression, which together formed a typical "strike package" of 12 to 24 aircraft. In the 1991 Gulf War, two cruise missiles with a range comparable to that of an F-4 (without refueling) could accomplish the same job for about $2 million. In 1972, more than a million pounds of aircraft left the runway at the beginning of a mission; 25 years later only 4,000 pounds were needed, a 25-fold reduction in gross weight and a 10-fold decrease in cost, not to mention the value of the lives that were saved.

Uninhabited Air Vehicles
These dramatic decreases in vehicle size and mission cost were principally enabled by the microelectronics revolution, which reduced avionics mass while providing real-time computation, navigation, electro-optics, and autonomy that greatly improved the accuracy of weapons and so reduced the mass of the weapons required. Perhaps most important, microelectronics have enabled the elimination of people from aircraft, thereby removing an important limitation to shrinking aircraft size.

Uninhabited (and largely autonomous) air vehicles (UAVs) are the latest innovation in military aircraft. Although the concept of UAVs, and examples of limited utility, have been around for more than 50 years, advanced avionics have enabled highly capable UAVs only in the last decade. The value of the UAVs used in recent conflicts has been acknowledged by the operational community. The largest of the current vehicles is Global Hawk, a high-flying reconnaissance aircraft weighing about 25,000 pounds with the wingspan of a small airliner. Global Hawk has a demonstrated transpacific range and endurances of longer than 24 hours. Its smaller and perhaps better known cousin, Predator, is the size of a light plane. In the war in Afghanistan, Predator became the first UAV to fire weapons in combat. The concept of uninhabited fighter and bomber aircraft is being advanced in flight testing of the X-45 and X-46 uninhabited combat air vehicles (UCAVs).

In the Iraq conflict, a host of smaller UAVs have been fielded, some weighing as little as 5 to 10 pounds; these include Pointer, Dragon Eye, and Desert Hawk. These small UAVs are flown by teams of a few soldiers for local reconnaissance or base security surveillance. As a historical footnote, these model airplane-sized UAVs sell for about $25,000, the same price the Wright brothers charged the U.S. government for its first airplane.

Jane’s All the World’s Aircraft is an 800-page annual compendium of airplanes currently produced and used around the world. The popularity of UAVs can be judged by the size of Janes’ Unmanned Aerial Vehicles and Targets, which is just as thick. The primary attraction of smaller UAVs is their low cost compared to the cost of manned aircraft. For this reason, UAVs are being embraced by operators, even though they may have less capability than the more expensive manned systems they replace.

The low price of UAVs may eventually alter the business landscape of military aviation. Because of the low development cost of small UAVs, new, small companies can and are entering the market, resulting in a proliferation of offerings. In contrast, current large aerospace concerns are organized to produce $100-million airplanes in $100-billion programs. If large manned programs are displaced by low-cost UAVs, large companies will find themselves challenged to establish viable business models for systems that cost only a few tens of thousands of dollars. Military combat pilots also feel threatened as it seems increasingly likely that no new manned combat aircraft will be in production by midcentury.

As military aviation transitions to and gains experience with UAV operations, it is also increasingly plausible that unpiloted flight will move into the commercial air transport arena. Routine transport flights are no more challenging than combat missions in terms of decision making and autonomy, so that it is reasonable to extrapolate that in the not too distant future, routine commercial flights could be automated. But what about nonroutine events and emergencies? It is true that pilots with extraordinary skills have saved severely damaged aircraft. But less skilled pilots are responsible for the majority of aviation accidents.

In the history of commercial jet aviation, about 70 percent of accidents have involved crew error. As aircraft have become more mechanically reliable, one category of error, controlled flight into terrain (the pilot literally flying an airworthy airplane into the ground because of a lack of situational awareness) has become the leading single cause of fatal accidents. Crew error implies that appropriate, established procedures were not followed. Thus, automation may not have to be as flexible as the very best pilots in all possible situations to maintain current standards of air safety, or even to improve upon them. Automated controls that do nothing more (or less) than reliably "follow the book" may result in a safer air fleet on average than we have now.

Technical issues aside, civil aviation may be harder to change than military aviation. Labor is a major issue in civil aviation, and it is difficult to imagine that pilot unions will look kindly on automated transports. Perhaps only new organizations that do not have such stakeholders will be able to deploy the new technology.

Once automated transport airplanes are ready to go, will anyone dare to board them? Probably not without a significant experience base in air cargo or military operations and strong additional inducements, such as low fares. Indeed, the last few decades of air travel suggest that the traveling public values low fares above all else.

At the start of the second century of flight, it is appropriate to ask how large an airplane need be. Aircraft exist mainly for transportation - of people, freight, sensors, ordinance, etc. We can classify missions as either mass specific (carrying things by the pound, such as freight or people) or function specific (accomplishing a task, such as reconnaissance, air superiority, or ground attack). The payload for a mass-specific mission is fixed by assumption (although in many cases the payload could be divided among several vehicles if that proved advantageous). In contrast, the payload mass for a function-specific mission is not fixed. It is determined by payload physics, contemporary technology, and the level of investment.

How small can a payload be? Take the example of destroying a bridge. Given the accuracy of current weapons (10 meters or less), perhaps 500 to 1,000 kilograms of explosives are required. But the minimum explosive mass required to destroy a bridge might be the mass carried by a human sapper who places a few charges at key locations, no more than 10 to 100 kilograms. With advances in explosives and warhead design, the amount might be reduced even further. A sufficiently capable group of future small air vehicles could deliver their payload right to the critical spots, flying under the bridge if necessary, and accomplish the mission with one-tenth to one-hundredth the total payload required by current guided weapons. This implies that the vehicles might be an order of magnitude or more smaller than today’s guided weapons for many targets.

Another example is a visual reconnaissance mission. Thirty years ago a payload consisting of a television camera and microwave downlink weighed about 100 kilograms and was carried on a remotely piloted vehicle with a takeoff weight of 1,000 kilograms. That same payload functionality can be realized today in a gram or two. Thus, the vehicle required to carry a payload that can accomplish the same mission could weigh no more than 50 to 100 grams.

Microair Vehicles
Based on this idea, the Defense Advanced Research Projects Agency (DARPA) initiated a microair vehicle (MAV) program in the late 1990s. DARPA somewhat arbitrarily defined an MAV as an air vehicle measuring 6 inches or less in every dimension. The concept is traceable to a Rand Corporation workshop in 1992 (Hundley and Gritton, 1992) that was developed in depth at the Massachusetts Institute of Technology Lincoln Laboratory (Davis et al., 1996). One of the most successful aircraft developed in this program was the AeroVironment Black Widow shown in Figure 3 (Grasmeyer and Keenon, 2001). With a 6-inch wing span and a 56-gram mass, it can fly for nearly 30 minutes on high-performance batteries and broadcast color video images from a distance of 2 kilometers. The Black Widow was a capable flyer, but it was subject to a hazard peculiar to very small, quiet air vehicles - bird attacks. For large aircraft, such problems have largely been limited to 1950s horror movies.

The DARPA program was a great learning experience for both the technical and operational communities. The technical community discovered that the 6-inch size was right at the edge of the state of the art at the time. Some aspects of the 6-inch airplane came out as expected. Aerodynamics at this scale could be readily calculated so that the poor performance relative to large airplanes (lift-to-drag ratios of 3 to 6 rather than 15 to 20) was no surprise. Instruments for navigation and control (such as 6-gram GPS receivers and 1-gram gyroscopes) were commercially available.

Designing subsystems that did not exist at this scale was more difficult. Propulsion and power systems were two of the most vexing problems. The smallest model airplane engine is ten times too large for a 6-inch airplane that needs about 2 watts of flight power to cruise at 15 meters per second. Battery propulsion had the advantages of low noise and availability, but the best commercial batteries store 25 percent less energy per unit mass than the jet fuel used by a modern gas turbine engine. The energy storage problem was exacerbated by the relatively low volume of a small airplane. Poor aerodynamics combined with limited energy storage space resulted in short flight times, tens of minutes at best. Low transmit powers and small antennas limited radio communications to a range of 2 to 4 kilometers.

Stability and control systems proved to be challenging as well. These airplanes were first flown by remote control rather than autonomously, but the dynamics of aircraft at this size are too fast for most people to control. Fast dynamics can be tamed with control loops, but this requires a very small, very fast servo motor, much smaller than was then available on the market. Payload was another challenge. In the original Lincoln Laboratory study, the technological potential for payloads was in the range of a few grams; DARPA chose not to invest in payloads in its program. Thus, these early vehicles were limited to daylight imaging cameras. System integration was difficult because there was no room for components, such as electrical connectors, in an aircraft with a mass budget of only a few grams of avionics.

The Army and Marines were supportive of the MAV concept, although they wanted aircraft with longer endurances and ranges, day/night imaging, and the ability to fly and perch in urban environments - capabilities that first-generation MAVs could not provide. Also, although these vehicles may be disposable in wartime, they are still too expensive to lose in peacetime training. The principal conclusion in 2000 was that a 6-inch aircraft was too small for a vehicle with the desired performance characteristics. Thus, the near-term focus was shifted to the 8- to 16-inch size range, which has improved the aerodynamics and is better suited to existing payload and subsystem technologies.

What about in the longer term? One way to address the question is to look at the engineering disciplines underlying much of aviation - aerodynamics, propulsion, and structures. Figure 4 shows the trend in the size of air vehicles with payloads that are a constant fraction of the initial mass. The calculations for this figure are for subsonic aircraft optimally designed for maximum range at each size (Drela et al., 2003). The figure also shows the performance of the Global Hawk and TAM-5, which have many design requirements in addition to long range, especially low cost. The structure becomes more efficient as aircraft size decreases, but the aerodynamics and propulsion become less efficient. The cubed-square law results in the volume decreasing relative to the area as size is reduced. Thus, smaller aircraft (less than a few thousand kilograms) have somewhat inferior aerodynamics and propulsion and relatively less fuel volume, and therefore less range. Nevertheless, aircraft that weigh less than a kilogram can have very useful, even transoceanic range. This should come as no surprise because birds as small as a 6-gram hummingbird migrate more than 2,000 kilometers without refueling.

As airplanes are scaled down, the aerodynamic and engine performance, the volume/wetted area, and the communications range all decrease while the necessary control bandwidth increases. However, some aircraft parameters do not vary with size. Flight speed, for example, is a function of installed power, not size; flight altitude is a function of flight speed. Recent transatlantic model airplanes had cruise speeds and altitudes comparable to those of the 1919 NC-4 because the performance of current model airplane engines is only marginally better than the performance of the large engines of 80 years ago.

Takeoff and landing speed and, therefore, runway length scale with cruise speed and installed power, rather than aircraft size. This implies that 6-inch transonic aircraft would require runways comparable in length to those of large aircraft (1 to 2 miles). Of course, the runways don’t have to be very wide. A more practical solution may be based on favorable structural scaling with decreasing size, which implies that variable geometry is much less costly for very small aircraft than for very large ones. Birds make good use of variable geometry when landing and taking off.

A major reason for the relatively low performance of current MAVs is the poor performance of subsystems, especially the propulsion subsystem. Historically, small airplanes have had small budgets to solve large engineering challenges, and progress is paced by the level of investment. The recent development of semiconductor-based micromachined devices, known as microelectromechanical systems (MEMS), has opened the way to new approaches to small engines. A micromachine gas turbine engine, for example, is sized to power an MAV in the 50-to 100-gram class (Figure 5) (Epstein, 2003). The performance of early versions will be no better than the performances of early turbojets in the 1940s, but greatly superior to the performance of battery-powered vehicles. These engines should increase flight speed and extend the range of MAVs. Other approaches, such as high-performance, miniature, internal combustion engines and fuel cells, are also being pursued. MEMS, improved microelectronics, and associated technologies can greatly improve air vehicles, down to the scale of a few inches.

Can aircraft be smaller still? Most flying insects are an order of magnitude smaller than MAVs, but there is comparatively little quantitative analysis or engineering experience at these length scales. Developments in biotechnology and nanotechnology may help, but, at the moment, neither the utility nor the challenges of subcentimeter-span aircraft are well understood. One thing is clear though, smaller is more difficult.

The Future
Forecasting the future of aeronautics can be risky. History is littered with inaccurate predictions of the future of flight. Lord Kelvin once opined that "aircraft flight is impossible." Millikan, von K?rm?n, Kettering, and others stated in a 1941 National Academy report that "the gas turbine can hardly be considered a feasible application to airplanes." Unbeknownst to them, the first jet plane had flown in Germany the previous year. Nevertheless, it seems fitting that I close with some remarks about what lies ahead.

First, autonomous air vehicles of all sizes will predominate, especially in military aviation, the inevitable result of the continued development of microelectronics and software. Second, the performance of UAVs will improve rapidly both because of increased investment and because small systems can be developed much more quickly than large military aircraft, which now take decades to reach the field. Indeed, progress can be faster for small aircraft for much the same reason that geneticists study fruit flies rather than elephants. Third, improving technology will enable the development of very capable air vehicles as small as a few inches in size. Large aircraft will always have their place, but the future of aeronautics will be small.

Brzezinski, M. 2003. The Unmanned Army. New York Times, April 20, 2003.
Davis, W.R., B.B. Kosicki, D.M. Boroson, and D.F. Kostishack. 1996. Micro air vehicles for optical surveillance. Lincoln Laboratory Journal 9(2): 197-213.
Drela, M., J.M. Protz, and A.H. Epstein. 2003. The role of size in the future of aeronautics. Paper no. AIAA-2003-2902. Reston, Va.: American Institute of Aeronautics and Astronautics.
Epstein, A.H. 2003. Millimeter-Scale, MEMS Gas Turbine Engines. ASME Paper GT-2003-38866. In Proceedings of ASME Turbo Expo 2003. New York: ASME.
Grasmeyer, J.M., and M.T. Keenon. 2001. Development of the black widow micro air vehicle. Paper no. AIAA-2001-0217. Reston, Va.: American Institute of Aeronautics and Astronautics.
Hundley, R.O., and E.C. Gritton. 1992. Future technology-driven revolutions in military operations: results of a workshop. Santa Monica, Calif.: RAND National Defense Research Institute.

See PDF version for figures.

About the Author:Alan H. Epstein is R.C. Maclaurin Professor of Aeronautics and Astronautics, Massachusetts Institute of Technology, and an NAE member.