GPS Declassified

by Richard D. Easton

  In 1959 Anderson discussed the use of range (distance) and range difference measurements for satellite navigation systems with Navy captain Alton W. Moody of NASA, who was president of the Institute of Navigation from 1959 to 1960. Range measurement computes the distance a radio wave travels from a satellite to a receiver, basing its calculations on the travel time, since radio waves move at the speed of light. Seventeenth-century astronomers noticed that observations of the Jovian moons were affected by whether the earth was on the same side of the sun as Jupiter or on the opposite side. This phenomenon showed that light travels at a finite speed. Nineteenth-century physicists hypothesized that it travels through a medium called ether. This reflected a philosophic aversion to the concept that space is largely a vacuum. Physicist Albert Michelson and chemist Edward Morely proved that the speed of light is a constant in a famous series of experiments in 1887. This disproof of the ether theory helped inspire Albert Einstein to formulate the special theory of relativity. Light travels at 186,282 miles per second. Consequently, the receiver is 18,628 miles from the satellite if the signal takes a tenth of a second to travel from it. Thus, one knows that the receiver is somewhere on the surface of a sphere 18,628 miles from the satellite.

  Anderson submitted a satellite navigation proposal in January 1963 to NASA, which awarded him a contract. He proposed two modes of operation: an active surveillance mode for transoceanic air traffic control and maritime safety and a passive navigation mode that would enable an unlimited number of users to determine their position from the satellite signals. There was more interest in the surveillance mode, since it was advantageous to know the position of commercial aircraft. For military applications, a passive mode is essential. In 1966 and 1967 NASA launched the ATS-1 (Applications Technology Satellite) and ATS-3 satellites, operating in the aircraft VHF band, which successfully tested Anderson’s approach.4 In 1964 Anderson proposed a navigation system with twenty-four satellites in four orbital planes with six-hour orbits at a fifty-one-degree inclination.5 Ground stations would send time signals to the satellites, which then would relay them back down to receivers; thus, Anderson proposed a repeater system. Placing atomic clocks in the satellites did not appear feasible in the foreseeable future. Based on available evidence, Anderson was the first to propose a twenty-four satellite navigation system in a midaltitude configuration. These are important aspects of GPS today, and the selection of a passive design allowing unlimited users is a key reason for the system’s success.6

  Fig. 3.1. Roy E. Anderson discusses satellite ranging with Richard L. Frey, James R. Lewis, and Axel F. Brisken (left to right) at the General Electric Earth Station Laboratory in Schenectady, New York, circa 1971. (Courtesy Gladys M. Anderson)

  Harold Rosen, who designed the first geostationary communications satellite at the Hughes Aircraft Company, lured to his team Harvard-educated Don Williams, who had proposed a space-based navigation system. Formulated in 1959, it called for two satellites in geostationary orbit for a range-measurement navigation system. Hughes vice president and later president Allen Puckett was participating in a yacht race from the West Coast to Hawaii, and Williams thought his navigation system could give Puckett an advantage. However, Puckett did not know about the proposal, and Williams did no further work on the navigation system after he began working on Rosen’s communications satellite.7

  Fig. 3.2. Roger L. Easton, former head of the Space Applications Branch at the Naval Research Laboratory, a few years before he retired in 1980. (Courtesy Naval Research Laboratory)

  Roger Easton at the Naval Research Laboratory (NRL) joined the Rocket-Sonde Branch in 1952. This branch launched Viking sounding rockets—experimental research rockets that get their name from the nautical term for taking measurements—from White Sands Missile Range in New Mexico.8 The NRL’S Viking rockets should not be confused with the NASA project in the 1970s for a soft landing on Mars. Launching rockets in the 1950s was a chancy endeavor. Viking 8 in 1952, despite being bolted down, broke free and took off during what was supposed to be a static rocket test. Easton said that Milt Rosen, the head of the Viking project, looked as upset that day as any person he had ever seen. The 1955 Project Vanguard proposal that Easton cowrote with Rosen listed navigation as a beneficial use of satellites. It stated, “It would also be possible to determine the absolute longitudes and latitudes by observation of the satellite. Such observations would also yield the height of the observer above the center of the earth. ”9 After the NRL’S proposal prevailed over Wernher von Braun and the Army’s proposal as the official American satellite program in the International Geophysical Year, Easton designed Vanguard 1 and the Minitrack system to track satellites at the IGY-designated frequency of 108 megahertz. However, Sputnik transmitted at 20 and 40 megahertz, so tracking it required modifying Minitrack.

  From Tracking to Navigation

  After Sputnik’s launch, Soviet spy satellites became a concern. These satellites would be silent, not emitting a signal most of the time; consequently Minitrack could not use their signal for space tracking. The new system Easton proposed in January 1958, Space Surveillance (formally the Naval Space Surveillance System, or NAVSPASUR), was an important development not only for its role in tracking satellites over coming decades. It created a technical problem—keeping the system’s clocks synchronized—and Easton’s solution of putting the synchronizing clock in a satellite led directly to his proposal for a navigation satellite system.

  Space Surveillance worked as follows: “The system concept of NAVSPASUR is that of a continuous wave (CW) multistatic radar. A high-powered transmitter generates a large fan beam of energy, commonly called the ‘fence ,” which reflects signals from an orbiting object back to separate receiving stations. These receiving stations use large arrays of antennas as an interferometer to determine the angle and angle rates of arrival from the reflected signals. By observing the target satellite from several stations, the position can be determined; using multiple penetrations, the orbit can be inferred. ”10 Thus the transmitter sends a beam into space; if it hits a satellite, the beam is reflected to receiver stations east and west of the transmitter. Easton’s colleague, Martin Votaw, has called Space Surveillance “Minitrack with the transmitter on the ground rather than in the satellite. ”11 Space Surveillance required receivers about one hundred times the size of Minitrack’s.12 Smaller satellites or those at higher orbits are more difficult to detect. ARPA, the Advanced Research Projects Agency, was established in 1958 in response to Sputnik to prevent technological surprises. It approved funding for Space Surveillance in June 1958, and the initial satellite detection occurred in August 1958. This shows how rapidly the pace of technological change in space occurred after Sputnik. Navy captain David Holmes, who later played a major though unheralded role in GPS, pushed for installing a powerful, one-million-watt transmitter for Space Surveillance at Lake Kickapoo, Texas.13 In 1960 work began on a second picket, or fence, in Texas, which turned the system into a radar and allowed it to estimate satellite orbits in one pass.14 Space Surveillance transmitters and receivers were aligned in an east-west direction, whereas Minitrack receivers were aligned in a north-south direction. Early satellite launches aimed eastward to take advantage of the earth’s rotation. A north-south system was optimal to detect these satellites as their orbits crossed overhead. For spy satellites, a polar orbit is preferable, because each orbital track passes over a new swath of ground as the earth rotates beneath it. An east-west tracking system is optimal for detecting these satellites.

  Fig. 3.3. Map of the Space Surveillance System. A series of powerful transmitters across the southern United States beam signals in the shape of a fan into space. Signals reflect back to receivers when a satellite crosses the beam. (Courtesy Naval Research Laboratory)

  Fig. 3.4. Capt. David C. Holmes at the U.S. Naval Academy in 1943. (Courtesy U.S. Naval Academy)

  Peter Wilhelm, currently director of the Naval Center for Space Technology at NRL, worked in the Satellite Techniq
ues Branch under Martin Votaw when he joined NRL in 1959. Wilhelm helped build the first electronic intelligence satellite, named GRAB. His first project for Roger Easton was building satellites to calibrate Space Surveillance. They were given the name SURCAL. Wilhelm asked if there should be an off switch for the signal, but Easton did not think it was necessary. The launch of SURCAL 1 was unsuccessful, but SURCAL 2 was launched in December 1962, and it jammed the system for about ten seconds on each orbit as its powerful transmitter overwhelmed the Space Surveillance system. Wilhelm said that it was a big irritant until it burned up three years later.15

  The Timation (short for time navigation) program began in 1964. There is some uncertainty about the sequence of events that year. Synchronizing Space Surveillance’s clocks in transmitter and receiver stations was critical for precisely obtaining the distance to a satellite being tracked. The tracking process relies on measuring the time it takes radio signals to travel from the transmitter to a satellite and be reflected back to receiving stations. If the clocks are not synchronized, an error is introduced. Since radio signals travel at the speed of light—299,792,458 meters per second—an error of one millionth of a second (1 x 10-6) translates to an error of almost 300 meters. Easton recalled that in September 1964 he conceived using a clock in a satellite to synchronize the transmitter and receiver station clocks in Texas. In October he realized while on Naval Reserve duty that this would also make a good navigation system. However, other documents show that Timation’s origin was more complicated. In a May 1967 slide presentation, Roger Easton stated, “To my knowledge the idea of using passive ranging as a navigation technique was a result of a conversation between myself and Dr. Arnold Shostak of ONR [Office of Naval Research] in 1964. Dr. Shostak was explaining how the hydrogen maser worked and obtained its fantastic time-keeping ability. At the time I remarked that this device appeared to make passive ranging feasible. He agreed and I spent a week working on the idea. ”16

  He sketched more details in a June 9, 1964, memo showing how passive ranging works and discussing the accuracy of different types of atomic clocks. A July 1971 NRL chronology tracing the history of Timation states that it began in April 1964, so the conversation with Shostak probably took place then.17

  It is not surprising that a discussion about atomic clocks between two Navy scientists led to a space-based navigation system using precise clocks. As shown in the preceding chapter, there is a historical linkage between maritime navigation and precise timekeeping. The chronometer was also known as the marine chronometer. Prior to the advent of aircraft, maritime travel was much faster than land travel apart from railroads, which followed fixed routes. The first atomic clock was an ammonia maser device built in 1949 at the U.S. National Bureau of Standards. Office of Naval Research funding supported the development of the first cesium atomic clock in 1951 and of the first hydrogen maser atomic clock in 1960. A paper Easton coauthored in the 1970s mentions that “in 1963, Study Group VII of the International Radio Consultative Committee referred to the advantages which might be expected to accrue from the use of time signal emissions from artificial earth satellites, and urged studies of the technical factors involved. ”18

  Timation’s initial test on October 16, 1964, used an NRL engineer’s convertible. Matt Maloof drove along the unfinished Route 295 with a transmitter in his car, and people at NRL tracked his position.19 Chester Kleczek, an engineer at Naval Air Systems Command, commented that Maloof was impressed that they could tell when he changed lanes. The results were sufficient for Kleczek to convince his boss, John Yob, to approve $35,000 for additional work. This was the largest amount Yob could give on his own authorization. The bureaucracy might have held up a larger amount, and it could have drawn opposition from proponents of Transit, which was also funded by Naval Air Systems Command.20 The amount may seem small, even adjusted for inflation, but Easton commented that having an authorized program was important, and he could mix the seed money with other funds.21

  Kleczek was born in Boston and studied at the Massachusetts Institute of Technology, but the Depression forced him to drop out after two years for financial reasons. Later he completed his degree at Northeastern University. As a sponsor for Timation at Naval Air Systems Command, he faced arguments in Pentagon meetings such as: Why do they need another navigation system when the government already supports forty-three navigation systems? Kleczek responded by asking which of them could give a Navy pilot his position over the South Pacific. There was a system, inertial navigation developed by Charles Draper, that could do it, but Kleczek knew that the people he was speaking with were unaware of it. However, inertial navigation could not be reset in flight if the power failed. Kleczek won approval to go to production.22

  Kleczek feared that others would steal the idea for Timation and claim they invented it. Many people tried to find out how Timation worked, and Kleczek dodged their questions by telling them that NRL was still working on it. Kleczek recalled the period in a 2009 interview:

  I learned how APL worked. There was a satellite launch called Transit and they put out gravity gradient stabilization on a satellite and the damn satellite kept wiggling back and forth. They didn’t know why. And NRL built one with gravity gradient but they put a damper between the satellite and the gravity gradient satellite, so there wasn’t a strong coupling between the lower, the dummy, satellite and the real satellite and it’s sort of slowed it down this way and it worked, it stayed vertical. Well, APL wrote a letter and they backdated it [saying] they had already invented it. I said, “Oh, you think we’re going to tell them about Timation? ” [Laughing.] That’s what happened. But anyway that was just the skepticism that they would steal it and this was going on all over the place during those times.23

  After the convertible test, NRL scientists simulated satellites using airplanes. An important breakthrough in winning support for continued research was explaining passive ranging using satellites in comparison to celestial navigation. Navy officers had for centuries navigated by measuring angles to stars. The Timation technique transformed this into measuring the time a signal took to travel from a satellite to a receiver. This explanation made sense to the Navy and facilitated its acceptance.24

  Fig. 3.5. Transforming celestial navigation to satellite ranging. This diagram helped to explain how navigation based on time signals from satellites compared to the angular sightings used in celestial navigation. (Courtesy Naval Research Laboratory)

  NRL launched the Timation I satellite in May 1967 and conducted initial tests internally. Researchers gave Pentagon officials the first outside demonstration at the John Ericsson statue near the Lincoln Memorial in October of that year. They selected the location for several reasons. First, good sightlines were important, given that the satellite would be visible for only about thirteen minutes (it was in orbit about seven hundred miles above the surface of the earth). Second, Ericsson’s invention of the screw propeller and the USS Monitor, the first ironclad battleship, made this an appropriate setting for demonstrating a major advance in navigation. Third, it was close to both NRL and the Pentagon. The successful demonstration yielded further funding. An article published to mark the fortieth anniversary of the test stated,

  The plan was to take one measurement at Time of Closest Approach [TCA] and one on either side between TCA and the horizon. As the data streamed in, [Don] Lynch called out which set of side tone readings should be resolved. Alick Frank then read the chart recorder deflections to James [Buisson, an NRL physicist], who resolved them and passed the result back to Lynch to be plotted on the intercept chart on the makeshift table. “The three points that Don chose were beautiful. They intersected very closely ,” Buisson reminisced later. After thirteen action-packed minutes, the satellite sank below the horizon and the pens fell silent.25

  Easton wrote in the May 1967 slide show referenced earlier that constellation studies were deferred due to budget limitations but that both geosynchronous and midaltitude (mainly eight- and twelve-hour) circu
lar polar orbits were being considered. Other tests were done with cars and boats, sometimes with unexpected occurrences. James Buisson and other NRL scientists were in a government van in Virginia, stopped on the side of the road waiting for the Timation I satellite to rise so they could track it. Buisson recalled, “A policeman stopped and asked us what we were doing. After we gave him a very detailed and long explanation about our atomic clock and the satellite, etc., he gave up trying to understand and said something like, ‘Good luck on your experiment ,” and he drove off. ”26

  Different Needs, Different Systems

  Rival navigation systems were proposed in the 1960s. The Joint Chiefs of Staff’s Navigation Study Panel specified that navigation systems provide three-dimensional instantaneous position fixes worldwide within a specified accuracy. Accuracy in a satellite navigation system depends on many factors, including the accuracy of individual clocks, the synchronization between them, precise knowledge of each satellite’s orbit, and corrections for changes in the signals caused by ionospheric distortion. Comparing two signals with different frequencies permits calculating and correcting for ionospheric distortion, so Timation II, launched in 1969, broadcast on two frequencies. NRL scientists extensively studied possible constellation configurations for Timation and determined that twenty-seven satellites, nine each in three evenly spaced orbital planes, with eight-hour orbits would be optimal. Ground stations were to be placed in Alaska, St. Croix in the Virgin Islands, Guam, and Samoa.27 Security concerns limited ground locations to either the United States or U.S. territories. This restriction was eliminated in the 1990s after the end of the Cold War with the placement of stations in less secure areas, such as Diego Garcia in the central Indian Ocean. Eight-hour orbits were preferred to twelve-hour orbits since they provided more frequent clock updates and minimized the problems from not having a station in the Indian Ocean. Twelve-hour orbits became more feasible when atomic clocks replaced quartz crystal oscillators in the satellites, reducing the need for ground updates.