(Extracted from Unmanned Aviation: A Brief History of Unmanned Aerial Vehicles, written by Laurence R. "Nuke" Newcome and published by AIAA in 2004.)

     Unmanned and manned aviation originated in the same era, and a number of the early pioneers of flight, including Orville Wright and Glen Curtiss, contributed vigorously to the development of both. World War I was responsible for encouraging the growth of both manned and unmanned aviation, but certainly not in equal proportions. Whereas manned aircraft progressed from a few hundred machines flying stunts in a few countries in 1914 to tens of thousands flying with a military purpose worldwide in 1918, unmanned aircraft had just begun to move from the lab bench to limited production over those same years. The British pursued radio controlled aircraft, the Germans wire-guided ones, and the Americans ‘programmed’ ones.
     Developments in the manned and unmanned fields diverged even more rapidly following World War I for a variety of reasons. The single largest of these reasons was insufficient technology. The development of unmanned aircraft hinged on the confluence of three critical technologies, in addition to that of flight itself: automatic stabilization, remote control, and autonomous navigation. Elmer Ambrose Sperry was the first person to attempt to address all three in a single unmanned aircraft design.

     Born in 1860, Sperry became a first-rank inventor in an age of great inventors who also possessed a keen business sense. His early work involved arc lamps and this may have brought him into contact with Peter Cooper Hewitt, a contemporary inventor of electrical lighting and fellow New Yorker. Neither Hewitt nor Sperry evidenced any interest in aviation until Sperry's work with gyroscopes for stabilizing ships led him to attempt the development of a gyrostabilizer for airplanes in 1909. While Sperry's intent was to improve the safety of flight by providing a mechanical sense of wings level to a pilot with vertigo or disorientation, in doing so, he also solved a key hurdle to unmanned flight: Stabilized flight in the absence of a pilot's inputs. But his 30-pound gyrostabilizer, besides being excessively heavy, performed poorly when it encountered the three dimensions of flight.
     Receiving encouragement from aviation pioneer Glenn Curtiss, Sperry reattacked the problem in 1911, using smaller gyros in the airplane's pitch, roll, and yaw axes and coupling them to the aircraft's controls by servomotors. Curtiss helped pique the U.S. Navy's interest in this development, and they dispatched Lieutenant Gordon Ellyson to Curtiss' plant in Harnrnondsport, New York, in 1912 to test fly the device. Sperry continued improving the device while Curtiss tried to interest the U.S. Army in it also. But after installing it in an Army aircraft in San Diego, Curtiss had several crashes, and Army interest quickly waned.
     In 1915, the Navy appointed both Sperry and Hewitt to its Naval Consulting Board, and Sperry, as chairman of the Board's Mines and Torpedoes Committee and a member (along with Hewitt) of its Aeronautics Committee, had little trouble gaining Navy approval and funding that October to continue developing what had come to be called an "aerial torpedo." The aerial torpedo program was broken into two phases, the first to develop a gyrostabilized, bomb-carrying drone with a distance gear (i.e., a preprogrammed version), followed by the addition of radio controls for directing the torpedo from an accompanying airplane (i.e., a controllable version). The Navy's concept of operations was to employ the torpedoes against German U-boat bases and munitions factories from distances of up to 100 miles. Flight tests began in 1916 using Curtiss N-9 seaplanes with safety pilots in them to perform the takeoffs and landings, and by November, 30-mile flights with accuracy errors of 2 miles (400 feet per mile flown) were being achieved.
     Sperry next ordered six Speed Scout airframes from Curtiss in October 1917, and the first was delivered just within the contract's 30-day deadline. This design was the first purpose-built unmanned aircraft. It had an empty weight of 500 pounds, a range of 50 miles, a top speed of 90 miles per hour, and was to carry a payload (bomb) of 1000 pounds. The record of the twelve flights made with these six UAVs before they were all expended reflect failures due to combinations of inadequate knowledge of the aircraft's stability characteristics and gyrostabilizer failings, but mainly due to underestimating the difficulty in getting an unmanned aircraft off the ground. The autopilot that had worked so consistently on the N-9 trials proved incapable of stabilizing the aerial torpedo after catapulting due to precession from the sudden acceleration.

Table 3-1. Summary of Curtiss-Sperry Aerial Torpedo Flights.

     In this list of failures however sits a singular accomplishment, that of the first successful flight by a powered unmanned aircraft, unmanned aviation's counterpart to the Wright brothers flight 14 years earlier. On 6 March 1918, a Curtiss Sperry Aerial Torpedo catapulted cleanly into the air, flew its planned 1000-yard flight, then dived at its preset distance into the water off Copiague, Long Island. True to the definition of a UAV, it was recovered and later reflown.

     Major General George O. Squire, Chief of the Army's Signal Corps in 1917, who had witnessed one of Sperry's successful demonstrations to the Navy, directed a similar Army effort be created in secret. In January 1918, the Army awarded a contract to Charles "Boss" Kettering and the Dayton Wright Airplane Company to develop and produce 25 Liberty Eagle aerial torpedoes, or as they became known as, "Kettering Bugs." Half the size of the Curtiss Sperry Aerial Torpedo, the pasteboard Bug was built to navigate to a target 50 miles away using preset mechanisms (including pneumatic values from his wife's player piano), then short out its engine's ignition and dive with its 200-pound bomb on to its target. Orville Wright, Kettering's technical consultant, added dihedral to the Bug's wings to improve its gust response. On its first successful test flight, the Bug wandered away from Dayton where it was launched and was hotly pursued by several cars led by then-Colonel Hap Arnold until it crashed, an hour later, into a farmer's barn near Yellow Springs, Ohio.

     The capability to remotely control the flight of an unmanned aircraft by radio was pursued by both Lawrence Sperry and the Army using Sperry's midget Messenger biplane and by Carl Norden (of World War II bombsight fame) and the Navy with a Curtiss N-9 seaplane. The Army effort ended with Sperry's death in an unrelated crash in 1923. But after several years of testing with safety pilots always on board, on 15 September 1924 the Navy launched the N-9 without a pilot, flew it for 40 minutes during which it executed 50 commands, then landed it successfully, replicating a similar British feat 12 days earlier with the RAE 1921 Target. In both cases, a robotic aircraft had taken off, flown through a series of maneuvers, and landed, all under remote control. During these same years in Britain, the Royal Aircraft Establishment was also developing a cruise missile, working up to flight ranges of 300 miles before beginning tests with live warheads from a RAF station near Basrah in present-day Iraq.

     Target drones were introduced in the 1930s, first in England then in the U.S., as a spin-off from these early cruise missile efforts. By the end of the decade, hundreds were regularly being flown in both countries to train their anti-aircraft gunners. In the U.S. Army, target drones sprang from the new hobby of RC model planes, introduced by Reginald Denny, who ran a Hollywood hobby store as a sideline to his acting career. Denny modified his model plane design to meet the Army's training needs and sold over 15,000 of them to the Army before and during World War II. Some of them were assembled by 19-year-old Norma Jean Dougherty, a Denny employee who went on to a successful Hollywood career herself under the name of Marilyn Monroe. Concurrent development of target drones for the U.S. Navy, under Commander Delmar Fahrney, led to the establishment of Utility Squadron Five, the first drone squadron, in March 1941 at Cape May, New Jersey. Fahrney's success inspired the Navy to attempt television-guided and radar-homing efforts with warheads during the war, which led to the Gorgon PGM.

     Despite over two decades of work developing the technology for unmanned flight in the U.S. and Britain, it was Germany that deployed the first operational cruise missiles and PGMs. Because the earlier work of Sperry and Kettering on 'aerial torpedoes' was kept under wraps between the two world wars, the Fieseler Fi 103, or V-1 "buzz bomb," introduced the public to cruise missiles and in a larger sense to robotic warfare. From its first use against England on 12 June 1944 to its last on 3 March 1945, some 10,500 V-1s were launched from coastal ramps or from these bombers, with just over 2,400 reaching their targets, predominantly in London. Sixteen percent of the V-1s were lost due to mechanical failure. Nearly 4000 were destroyed by British fighters, artillery, or barrage balloons, a 38 percent loss rate that justified the German use of unmanned aircraft rather than having risked their aircrews over Britain at that point in the war. Conversely, 2,900 Allied aircrew members were lost defending against the V-1 attacks. A 1944 British assessment comparing the cost to the Germans of waging the V-1 campaign against the cost of its impact on the Allies concluded the V-1 offered a 4 to 1 return on its investment, thus cementing the role of cruise missiles in future wars (Armitage, 1988). German successes with PGMs were fewer but no less dramatic. A single radio-guided Fritz glide bomb nearly sank the U.S. cruiser Savannah in the Mediterranean Sea in September 1943.

     Reconnaissance drones burst on to the military scene in the 1950s. In 1952, Northrop bought Denny's small company, added cameras to his latest target drone variant in 1955, and sold it to the Army as the SD-1 Observer (later redesignated as the MQM-57), the first tactical reconnaissance UAV. Between 1959 and 1966, the Army bought 1,455 Observers, and the model spread to use in other NATO countries. The Army also began developing the SD-2 Overseer and SD-3 Sky Spy tactical UAVs and the SD-4 Swallow and SD-5 Osprey long range, strategic reconnaissance UAVs, but all were cancelled before reaching a production decision. The Air Force explored, then abandoned, a reconnaissance version of its GAM-67 Crossbow cruise missile between 1954 and 1957. And the Marine Corps tested its two-man Bikini UAV for small units, forerunner to its deployment of Pointer and later Dragon Eye mini-UAVs, in the 1960s. By the Cuban Missile Crisis, the Air Force had modified a small number of its Ryan Firebee target drones to carry cameras and return their film, a capability which was used extensively (over 3500 sorties flown) during the Vietnam conflict.

     The 1950s also saw the maturation of inertial navigation systems, the third and final key technology to enabling true unmanned flight. Charles Stark Draper of MIT led the effort to develop this capability to provide self-contained navigation and guidance from the end of World War II right up to its use on the Apollo moon flights. Today, INSes updated by GPS signals provide the precision in PGMs and ensure the return of UAVs.

     Contrary to current belief, the recent use of Hellfire missiles from Predators over Afghanistan and Iraq did not introduce the strike role to UAVs. That was done when the Navy's Gyrodyne QH-50 drone helicopter demonstrated it could carry and employ anti-submarine torpedoes in 1962. This same UAV subsequently carried a minigun and dropped an assortment of small munitions in covert operations over Vietnam in the late 1960s. In 1972, the Air Force's Have Lemon program demonstrated the delivery of Maverick and Stubby Hobo missiles from unmanned Firebee UAVs for potential use in the suppression of enemy air defenses (SEAD) role. However, the end of the war in Vietnam quickly ended these excursions into the strike role by unmanned aircraft. Interest in unmanned combat aircraft was subtly rekindled by the 1980's HiMat and the 1990's X-36 technology demonstrators, which led to the X-45 effort in the late 1990s and the current DARPA Joint Unmanned Combat Aircraft System (J-UCAS) program.

     As it approaches its own centennial, unmanned aviation can point to an impressive list of accomplishments:

- 1918 First flight by a powered unmanned aircraft (Curtiss Sperry      Aerial Torpedo)
- 1923 First radio controlled flight (RAE 1921 Target)
- 1933 First UAV used in the target drone role (Fairey Queen)
- 1943 First use of a PGM in combat (Fritz)
- 1944 First use of a cruise missile in combat (Fi 103 "V-1")
- 1946 First UAV used in the science research role (Northrop QP-61)
- 1959 First UAV used in the reconnaissance role (Northrop SD-1      Observer)
- 1960 First untethered flight by an unmanned helicopter
      (Gyrodyne QH-50)
- 1962 First UAV used in the strike role (Gyrodyne QH-50)
- 1998 First trans-Atlantic crossing by an unmanned aircraft      (Aerosonde)
- 2001 First trans-Pacific crossing by an unmanned aircraft (Global      Hawk)

     Future UAVs will evolve from being robots operated at a distance to independent robots, able to self-actualize to perform a given task. This ability, autonomy, has many levels emerging by which it is defined, but ultimate autonomy will require capabilities analogous to those of the human brain by future UAV mission management computers. To achieve that level, machine processing will have to match that of the human brain in speed, memory, and quality of algorithms, or thinking patterns. Moore's Law predicts the speed of microprocessors will reach parity with the human brain around 2015. Others estimate the memory capacity of a PC will equal that of the human memory closer to 2030. As to when or how many lines of software code equate to "thinking" is still an open question, but it is noteworthy that pattern recognition by software today is generally inferior to that of a human.
     As for the controller of UAVs, he will eventually be linked to his remote charge through his own neuromuscular system. Today's ground station vans are already being superceded by wearable harnesses with joysticks and face visors allowing the wearer to "see" through the UAV sensor, regardless of where he faces or moves to. Special vests will soon provide him the same tactile sensations "felt" by the UAV when it turns or dives or encounters turbulence. Eventually, UAV pilots will be wired so that the electrical signals they send to their muscles will translate into instantaneous control inputs to the UAV. To paraphrase a popular saying, the future UAV pilot will transition from seeing the plane to being the plane.
     The push of unmanned flight has been the driving or contributing motivation behind many of the key advancements in aviation: the autopilot, the inertial navigation system, and data links, to name a few. Although UAV development was hobbled by technology insufficiencies through most of the 20th century, focused efforts in small, discrete military projects overcame the problems of automatic stabilization, remote control, and autonomous navigation. The last several decades have been spent improving the technologies supporting these capabilities largely through the integration of increasingly capable microprocessors in the flight and mission management computers flown on UAVs. The early part of the 21st century will see even more enhancements in UAVs as they continue their growth. The ongoing revolution in the biological sciences will next impact aviation and, together with future microprocessors, will enable intelligent, vice robotic, UAVs to fly over the Earth and other planets.