The NASA 737 Advanced Transport Operating Systems (ATOPS)
Program aircraft landing at Wallops Island in 1992.

Prior to World War II, aircraft approaches and landings in poor weather conditions were often highrisk endeavors. Early civilian air navigation systems provided only basic lateral position information, and 1930s airline pilots flying in low visibility conditions had to take a heading off a radio navigation station and then use speed, time and distance calculations to figure out when they should see the runway at their destination. Safe descents were dependent on the accuracy of their calculations and all final approaches had to be visual, because there was no instrument guidance system that could direct a pilot on a safe, precise descent profile clear of terrain.

The NASA 737 research aircraft on the Wallops runway in 1987 with the Microwave Landing System equipment in the foreground.

The successful demonstration of a full Instrument Landing System (ILS) in 1937,[Ref 4-1] therefore, marked a significant advancement for air transport operations. For the first time, pilots could tune in a radio signal emanating from an airport runway and receive precise lateral and vertical guidance on an approach and landing. The ILS broadcast a straight, narrow VHF/UHF signal beam that started at the runway and rose at a steady, three degree angle. By centering the vertical and horizontal needles on his ILS instrument, a pilot could fly down the ILS "beam" to the airport, descending on a gentle, three degree glide slope that was lined up perfectly with the runway centerline and clear of any terrain or obstacles. The system not only improved the safety of airline operations, but also increased the reliability of air carrier service, since it allowed airliners to operate in a wider range of weather conditions.

U.S. airlines began using ILS approaches on a regular basis after World War II, and by 1949, the ILS had become the world standard for landing guidance systems. By the late 1960s, however, air traffic congestion and the need for more noisesensitive approach paths to airports had begun to demand a more capable and flexible landing system. One problem with ILS was that the VHF/UHF frequencies in which it operated had a limited number of channels. As air transportation became more popular, planners began to see a time when certain areas of the country would have more airports requesting instrument landing systems than the ILS frequency range could accommodate. The relatively low frequency range of the ILS was also susceptible to signal reflection, or "multipath," errors.

The Microwave Landing System approach vectors. The expanded coverage of the MLS tracked signals within an area 120 degrees wide and 20 degrees high from the runway centerline.

Another limitation of the ILS was that it allowed only one approach path to a runway. Airplanes would lock on to the ILS horizontal "localizer" and vertical "glide slope" up to 10 miles away from the runway, and fly a straight course in for a landing. Two of the top priorities identified by the Department of Transportation/NASA "CARD" study in 1971, however, were to increase the air traffic capacity of airports and to develop approaches that avoided noisesensitive areas.[Ref 4-2] The curved, segmented approach paths suggested by Jack Reeder and others in NASA and the FAA[Ref 4-3] could accomplish both of those goals, but they would require a more flexible landing guidance system than the ILS.

In 1968, Special Committee 117 of the Radio Technical Commission for Aeronautics (RTCA),[Ref 4-4] representing both military and civil U.S. airspace users, was formed to develop specific requirements and specifications for a new aircraft approach and landing system.[Ref 4-5] Four years later, the International Civil Aviation Organization (ICAO), an agency of the United Nations that oversees international civil aviation procedures and standards, began the official selection process for a new precision approach and landing guidance system standard. Member states were invited to submit system proposals, and ICAO's All Weather Operations Panel (AWOP) was given the responsibility of evaluating the proposals and making a recommendation to the organization as to which system should be adopted as the standard.[Ref 4-6]

By this time, the U.S. Federal Aviation Administration, working with the RTCA special committee, had already evaluated several types of technologies and had decided that a microwave frequencybased, scanning beam format would make the best landing system. The higher frequency microwave band would alleviate the frequency allocation and some of the multipath signal problems, and a scanning beam technique would provide broader signal coverage, allowing more flexible airplane approach paths. There were two ways of designing a scanning beam microwave system, however. One, called a "Doppler Scan" technique, used modulations in frequency to tell a pilot where he was in relation to the runway. The other, called the "Time Reference Scanning Beam" (TRSB) technique, used timereferenced sweeps of a single frequency to pinpoint an airplane's location.

In 1973, the FAA awarded feasibility and demonstration contracts for two separate industry teams to evaluate each of the two systems. ITT/Honeywell and Hazeltine/Sperry each developed a Doppler system, while Texas Instruments and Bendix Corporation each developed a TRSB system. The systems were evaluated by a 17man steering committee of the Microwave Landing System (MLS) central assessment group, made up of representatives from the Federal Aviation Administration, the Department of Transportation, the National Aeronautics and Space Administration and the Department of Defense. In early 1975, the TRSB microwave landing system was chosen as the U.S. candidate for international standardization.[Ref 4-7]

The performance and the cost of the TRSB and the Doppler designs were both actually very close, but the TRSB appeared to have slightly better performance at sites that were particularly susceptible to signal reflections, such as airports surrounded by high buildings or terrain. In the TRSB system, an airplane's horizontal and vertical location was determined by two separate microwave beams with very precisely timed scanning patterns. The first beam swept back and forth across the runway centerline (plus or minus 60 degrees on each side) at a rate of 13.5 times per second. The second swept up and down from the runway elevation to a position 20 degrees above that at a rate of 40 times per second. The airplane's horizontal and vertical position could be determined by measuring the time difference in between each signal transmission received by the aircraft. By integrating that data with distance information from a conventional distance measuring equipment (DME) transmitter on the airport, a receiver on board an airplane could accurately pinpoint the plane's location in relation to the runway. A third beam, scanning up and down over a 7.5 degree arc from the runway elevation 40 times a minute, could also be installed to provide flare guidance for automatic landings. The net result was a system that could provide precise manual or automatic landing guidance to an airplane anywhere in a wedgeshaped area that stretched120 degrees wide and 20 degrees high from the airport runway.[Ref 4-8

The TRSB system was not the only one submitted to ICAO, however. Australia also proposed a TRSB microwave system, but the United Kingdom was strongly advocating a Doppler Scan MLS. The Federal Republic of Germany proposed a dual DME system, and France submitted a plan based on a Ground Controlled Approach (GCA) technique. The U.S. representatives to ICAO had known that there would be other proposals. But after a couple of meetings, the FAA realized that the British were going to put up a very tough fight against the U.S. choice. To win acceptance for the U.S. candidate system, the FAA MLS program managers decided they needed to "look at all the assets available to us and see if we could exploit those in ways we hadn't exploited them before."[Ref 4-9] One of the assets they found was NASA's Boeing 737 Transport Systems Research Vehicle (TSRV).

Although the creators of the TCV program had not anticipated using the 737 for official demonstrations of the microwave landing system, they had hoped to test curved path and variable glide path angles with two demonstration MLS systems the FAA had installed at NASA's Wallops Island Flight Center Facility,[Ref 4-10] across the Chesapeake Bay from the Langley Research Center. In fact, several people in Langley's Flight Instrument Division (FID) had already begun to gather data from the microwave systems at Wallops Island before 1975. The researchers wanted to model the MLS signals in order to experiment with ways to process the information for the advanced aircraft displays and guidance systems they were developing. Langley did not have access to any actual airborne MLS receivers, so the data collection was conducted with a MLS receiver Langley engineers built out of offtheshelf high frequency receivers and high speed recorders and installed in a DeHavilland DH6C Twin Otter airplane. The makeshift receiver consisted of "a huge rack, three bays wide full of equipment," but it allowed the Langley researchers to get the MLS signal data they wanted.[Ref 4-11]

Several people in the antenna and microwave research branch at Langley had also begun experimenting with MLS aircraft antennas as early as 1974, evaluating designs and placement locations on an airplane that would create the broadest coverage and the least signal interference.[Ref 4-12] As a result, researchers at Langley already had some basic groundwork accomplished by the time the FAA asked Langley to support its efforts to sell the TRSB system to ICAO. This was fortunate, because the job the FAA wanted the 737 to perform was not an easy one.

The FAA's navigation division had originally asked the NASA researchers to provide 10 to 25 hours of MLS data collection flights on straight approach paths in the summer of 1975. [Ref 4-13] When the FAA MLS division realized how tough a fight the ICAO selection process would be, however, its managers decided that what they really needed was not just data, but a way to actually demonstrate some of the more impressive capabilities of the TRSB system. In the spring of 1975, the FAA asked the TCV program office if the TSRV 737 could perform a demonstration of curved path approaches and automatic landings to the ICAO All Weather Operations Panel. There was initially some concern within the TCV program office that the workload involved in preparing for a demonstration that complex would cause too many other research projects to suffer, but in July 1975, NASA agreed to participate in the demonstration.[Ref 4-14]

The task ahead of the Langley engineers was considerable. The demonstration was scheduled for May 1976, only 10 months away, and the 737 was not yet equipped to process MLS information or perform curved path approaches. In addition, the airplane was designed to use inertial navigation system (INS) data for autolands, which used internal gyros to determine movement from a known starting point. The FAA, however, wanted the ICAO demonstration autolands to be done without the use of an INS. [Ref 4-15]

Map of vectors flown for research with the Microwave Landing System flight tests in 1978. Demonstrates three different curved paths possible instead of the straight line path used for ILS.

There was not enough time to develop a new automatic control system that would operate off of MLS signals, so the engineers decided instead to convert the MLS information into a more conventional signal format the airplane's computers could recognize. A separate Singer Kerflot 2000 digital computer was installed to reformat the data from the MLS receivers and convey the translated information to the airplane's existing navigation and flight control computers. Curved approach paths that would line up precisely with the runway were designed into the navigation computer and the autoland system was adjusted to use MLS signals for its navigation and guidance cues, although it still needed acceleration information from a secondary source for smooth control of the airplane. [Ref 4-16]

The engineering and design challenges of creating all the various pieces of the system and integrating them together and with the other equipment in the airplane would have been formidable in any event. But this particular research endeavor had the added pressure of a very public and nonnegotiable deadline. If the system was not ready to go when it came time for the demonstration, it would reflect badly not only on NASA, but on the FAA and the United States itself. As a result, many people involved in the project put in extremely long hours throughout the fall and winter of 19751976.[Ref 4-17] The TCV team put the concepts, calculations and systems through three different levels of sophisticated simulations before they even put the equipment on the actual airplane. They then took the 737 up to the FAA's National Aviation Facilities Experimental Center (NAFEC) in March 1976 for more than five weeks of flight testing to work out any remaining problems in the equipment before the demonstration. [Ref 4-18]

Wind tunnel model tests of placement of the antennae for Microwave Landing System research in 1974.

Considering how quickly the system had been developed, the flight tests went well. They were not entirely without incident, however. Several days before the ICAO demonstration, for example, the flight crew was monitoring an automatic MLS guided landing in the airplane. Normally, when the flight path reached the point where the glide slope guidance began, the airplane would pitch down slightly as it began its final descent. Yet on this occasion, only 600 feet or so off the ground, the nose pitched up and kept rising. The safety pilots in the front cockpit took over, forced the nose down again and landed the airplane. Upon investigation, the researchers finally discovered that the problem was an incorrect algebraic sign on one of the computer tapes. Since the system was all experimental, the computer tapes were full of "patches" where programming modifications had been made. Every time the flight computers "crashed" (which happened often when the airplane hit a sharp bump on a runway), each patch would have to be reloaded individually. To try to simplify the process, a new tape had been made that incorporated all the patches, but a single algebraic sign had been transposed in the process, causing the pitchup problem. On another occasion, a disconnected computer lead caused the airplane to initiate a roll as it approached the runway on crosswind landings. There were problems with the MLS receiving equipment overheating and breaking down. The crew operated in horrendous wind and weather conditions throughout the month of March. But when the all weather operations panel of ICAO arrived for the demonstration, everything came together. [Ref 4-19]

The TCV program's demonstration at NAFEC consisted of repeated automatic flight demonstrations of two basic MLS approaches, followed by automatic landings. The first path was an "Sturn" approach, which incorporated a 90 degree turn to the right followed by a 180 degree turn to the left, rounding out onto a final approach leg 3 miles from the runway. The second path was a descending 130 degree turn to the runway with a straight, three mile long final approach. The demonstration lasted for 10 days, during which time 11 flights were completed. Each flight carried approximately 10 observers who rotated positions so that everyone got to see an automatic landing from both the aft flight deck, where the electronic displays were located, and the forward cockpit. Although the demonstration was conducted in severe wind and turbulence conditions, the results were impressive. The mean overshoot error of the airplane when turning final was about 30 feet, decreasing to 10 feet one mile before the runway. The mean vertical error at one mile was less than five feet. The performance of the second "EL2" elevation transmitter for flare guidance compared favorably with that of a radar altimeter, which was what airlines were using at the time for automatic ILS landings.[Ref 4-20]

Even more significant than the exact numbers in the research reports, however, was the fact that the ICAO panel members had not just observed a static viewgraph presentation, or even a flight demonstration of standard, straightin approaches using the new kind of guidance system. They had seen for themselves, with the help of the 737's vivid, visual electronic displays, a transport airplane perform maneuvers that no other transport airplane had ever been able to accomplish. The ICAO panel didn't have to imagine what the U.S. microwave landing system could do for air transport operations. They had already experienced it. [Ref 4-21]

Researchers in the TCV program continued to perform some additional experiments with the MLS at the New Jersey technical center during the summer of 1976, including manually flown, curved path approaches with even shorter final legs, but they thought their involvement in the MLS selection process was essentially finished. In fact, however, it was just beginning. [Ref 4-22]

One of the methods the FAA had used to evaluate the four candidates for the U.S. microwave landing system was computer simulation. A computer could model adverse effects that were not currently present at any airport, but were anticipated in the future as traffic and real estate development expanded. The agency had hired the Lincoln Laboratory, from Lexington, Massachusetts, to do the simulation work. So by the time the international evaluation of systems began, Lincoln was experienced in modelling MLS signals and diagnosing potential "multipath," or signal reflection, problems, and its work had a lot of credibility with ICAO members.

In July 1976, the Lincoln Laboratory presented a report to a technical working group of ICAO that indicated the Britishsupported Doppler Scan MLS might be more susceptible to multipath errors in some future situations than anyone had previously thought. The British modified the system's antenna scan pattern and alleviated the problem, but a new evaluation by Lincoln Laboratory indicated that the modification might make the system more susceptible to other multipath problems at certain sites. In March, 1977, the FAA asked Lincoln Laboratory to look more closely at one such potential location, in Brussels, Belgium. Using the official map of the Brussels airport, the lab ran a new simulation of standard and low aircraft approach paths and determined that there might, indeed, be a significant multipath problem with the Doppler MLS. The results were presented to a meeting of the all weather operations panel of ICAO in Montreal, Canada, eight days before its crucial vote on which landing system to recommend to ICAO for adoption as the international standard. The decision of the 10person panel was close, but the United Statesbacked TRSB microwave landing system was approved by a 6-4 vote.

Even before the AWOP vote, the British contingent promoting the Doppler system had been pushing for a competitive flight demonstration of the Doppler technique versus the TRSB system, arguing that regardless of what computer simulations showed, the reallife performance of the two systems was comparable. The British company that manufactured the Doppler system, Plessy Company, Ltd., also hired a lobbyist to promote the Doppler system to other elements of the U.S. government outside the FAA. The lobbyist conducted a campaign with Congressional representatives, the Department of Transportation, the White House, and the U.S. media to try to cast doubts on the TRSB system, and put pressure on the U.S. ICAO representative to delay the AWOP vote. These actions did not sit well with the FAA, of course, and by the time the Montreal vote was held, the debate between the two contingents had become heated.

The controversy took on new proportions immediately after the ICAO vote, however, when the British decided to test the Doppler system at the Brussels airport in order to prove the Lincoln Lab simulation wrong. Upon arriving at the airport, they discovered that one of the key buildings that had caused the multipath problem in the Lincoln Lab simulation did not, in fact, exist. It had been planned, so it was on the airport map, but it had never been built. When Lincoln Lab representatives travelled to Brussels and gathered their own onsite information about the airport configuration, and the lab reran the simulation, the results showed no significant difference in performance between the Doppler and TRSB systems.

Accusing the FAA of intentionally misleading Congress and ICAO and subverting the scientific process, the British lobbyist intensified his campaign to pressure the U.S into comparative flight tests of the two systems. He also implied that the reason the FAA did not want to conduct the tests was that they knew the demonstrations would show no difference between the two systems. The controversy was approaching the dimensions of an international incident, and after the issue reached the level of hearings before the House Government Operations government activities and transportation subcommittee in September 1977, FAA Administrator Langhorne M. Bond agreed to a series of international demonstrations.[Ref 4-23]

Once again, the aid of the TSRV 737 airplane was enlisted.[Ref 4-24] The first demonstration was held October 31 November 7, 1977 in Buenos Aires, Argentina, during a conference sponsored by the Organization of American States (OAS).[Ref 4-25] The demonstration flights were conducted at the Aeroparque Jorge Newbery, which had many approach path restrictions due to the fact that it was only 4 kilometers from the city center of Buenos Aires. The MLS equipment the FAA installed at the airport was more limited than that at the New Jersey technical facility, with an azimuth (lateral position) coverage of only 40 degrees on either side of the runway centerline. The FAA also did not install a second elevation transmitter for flare guidance, so the TSRV used a conventional radar altimeter for that portion of the autolands.

The TCV program researchers designed two descending, curved path approaches to the Buenos Aires airport that followed the Rio de la Plata river, minimizing the noise impact on the heavily populated areas under the long, straightin ILS approach. One of the paths had a 1.6 nautical mile final approach leg, while the other had only a 1.1 mile final. During the demonstration, a total of 56 automatic approaches and landings were made, with accuracies similar to those achieved during the NAFEC flights in May 1976. As important as the statistical results, however, was the fact that the demonstration allowed ICAO representatives from the different OAS countries to observe the U.S. system in action on board NASA's 737. Once again, the electronic flight displays gave the observers an impressive, visual illustration of the complex, automatic maneuvers that contributed greatly to the impact of the demonstration.[Ref 4-26]

Aeronautical chart of crowded New York airspace. The MLS would help prevent conflicts between airport flihgt areas.

Less than a month later, the NASA team took the airplane to the John F. Kennedy airport in New York City, where the U.S. TRSB system and the British Doppler MLS would both be tested. The New York metropolitan airspace was extremely congested, because there were three major airports (JFK, LaGuardia, and Newark, New Jersey) all located in close proximity to one another. As a result, some of the ILS approach paths for one airport overlapped the air traffic control zones of another. This was an area where curved path approaches could make a tremendous difference in airport safety and capacity.[Ref 4-27]

The demonstration flights lasted from December 5 13, and the weather conditions were harsh throughout the week. On five of the eight days, the approaches had to be flown with tailwinds exceeding 20 knots, and below freezing temperatures played havoc with the computer equipment on the airplane. On the first morning, for example, the electronic flight displays would not work. After an hour and a half of troubleshooting, the displays seemed to be fixed, but the same problem occurred again the next morning. The technicians in charge of the equipment finally figured out that the root of the trouble was simply that the display computers were getting too cold at night. They fixed the problem the first day by borrowing a hair dryer to warm up the affected computer component, and ended up taking that piece out of the airplane every night to store it in a heated trailer. In truth, the electronics technicians had one of the toughest jobs during the demonstrations, because in addition to flying on the airplane during the day, they had to fix any problems with the research equipment at night so the flights could be kept on schedule.[Ref 4-28]

The path the TSRV 737 demonstrated at JFK was the most demanding yet. It followed the "Canarsie" VOR approach to runway 13L, which was normally used only in visual flight conditions because the final leg was less than half a mile long.

The TSRV flew a total of 38 automatic approaches during the official demonstration at JFK and successfully completed 30 autolands. The safety pilots had to take over and land the airplane on eight approaches, but the performance of the airplane was considered "very successful," considering the adverse weather conditions and the fact that the .44 mile final leg left the autoland system very little time to capture the final approach segment.[Ref 4-29]

The British tested the Doppler system on the same runway two months later, but the British Aerospace (HS) 748 airplane had neither an autoland system or the capability to fly curved approach paths. So although the British airplane performed numerous automatic straightin approaches to a 50 foot decision height without any difficulty, the effect was not as dramatic as the 737's autoland performance on the difficult Canarsie approach.

The final MLS demonstration involving NASA's 737 was for ICAO's All Weather Operations Division, which had to approve or overrule the recommendation of the All Weather Operations Panel to adopt the TRSB design as the world standard for precision approach and landing guidance systems. The division meeting where the MLS vote would be taken was held in Montreal, Canada in April 1978, and both the British and the U.S. set up demonstrations of their respective systems.

The Microwave Landing System flight demonstrations at Montreal, Canada. The map depicts two curved-path MLS approaches demonstrated by the Langley 737 during the 1978 ICAO meeting.

The 737 demonstrated a curved path and an Sturn approach, designed to keep the airplane clear of an Indian reservation just south of the airport. One of the paths also had a glide slope angle of over four degrees, as opposed to the three degree descent angle used for ILS approaches. As in New Jersey, Argentina and New York, the FAA had observers ride on board the 737 so they could see for themselves, with the help of the aft cockpit's electronic flight displays, how well the airplane performed the curved, complex maneuvers and automatic landings. [Ref 4-30]

On April 19, 1978, the all weather operations division of ICAO voted 39 to 24 (with eight abstentions) in favor of the TRSB system. The decision remained politically charged up until the end, however, and the final vote was taken by secret ballot to protect the member countries from undue political pressure.

In truth, there was actually little technical difference between the British and U.S. systems. The TRSB design was a little more advanced, but the Doppler scan probably could have matched the TRSB performance with a little more development.[Ref 4-31] By using the TSRV 737 for the flight demonstrations, however, the FAA gained a couple of distinct advantages in the competition. First, it gave the U.S. system what Frank Frisbie called the credibility of "the NASA logo." Second, the fact that the TSRV airplane could not only perform complex automatic approach and landing maneuvers, but could also display them visually on electronic flight instruments, gave observers a dramatic and vivid impression of the TRSB system's capabilities that the British could not match.[Ref 4-32]

After the April 1978 ICAO decision, the organization began developing Standards and Recommended Practices (SARPs) for MLS. The target date for switching over to MLS as a primary landing guidance system was changed from 1985 to 1995, but the future of the TRSB microwave landing system seemed assured. The technology was proven, the ICAO world organization had given its stamp of approval, and after the SARPS were written, all that remained was to produce and install the equipment.

Yet 15 years later, only a handful of MLS installations had been completed, and it was uncertain whether the microwave landing system would ever be implemented, at least in the United States. In a telling example of how complex the technology application process can be, the MLS encountered so many delays and obstacles in the 10 years following the ICAO decision that it finally began to be overshadowed by an even newer technology: the Global Positioning System (GPS).

The FAA awarded the first production contract for 178 of the 1,250 planned microwave landing systems to the Hazeltine Corporation in Commack, New York, in January 1984. The specifications for the equipment required only a Category I level of performance, however, because the FAA planned to put the first systems at small airports which had only Category I ILS equipment, or no instrument landing capability at all, to help ease the transition to MLS.[Ref 4-33] Category I conditions were the least severe of the three lowvisibility condition categories, requiring at least 200 foot cloud ceilings and a half mile of visibility. Actually, the guidance signal of the MLS equipment did not change from Category I to Category III, but Category II and III installations required a much higher level of redundancy in the equipment.

The FAA planners hoped that by installing the first microwave landing systems at outlying airports, they would prove the flexibility of the system to provide coverage in areas unable to incorporate a standard ILS and begin to build a consensus of support for MLS among regional airlines and corporate aircraft operators. Interestingly enough, although the major airlines had initially pushed for the development of MLS in the late 1960s, they were less enthusiastic as the time came to implement the system. Some of the problems with the ILS had been rectified, and the growing financial concerns of the airlines following deregulation made them reluctant to purchase new avionics for their entire fleets unless they could see some immediate benefit from the new equipment. [Ref 4-34]

The problem with the FAA's strategy was that installing Category I equipment at outlying airports offered no opportunities to show the airlines any of the more impressive advantages the MLS could offer them, such as more efficient approach procedures or better automatic landing capabilities.[Ref 4-35] Nevertheless, had the 178 Category I systems been installed as planned, there might have been enough momentum to support the production and installation of the more capable Category II/III MLS equipment at airports used by the major airlines. But as it turned out, those first 178 systems were never delivered.

Hazeltine Corporation had promised it would begin delivering the MLS systems 18 months after its contract with the FAA was signed. The company ran into problems with software, however, and the delivery date was pushed back repeatedly. Four years later, Hazeltine had only delivered two systems, and the FAA finally cancelled the contract in 1989.[Ref 4-36] By this time, numerous questions were being raised about the necessity of MLS, not only by the airlines, but also by Congress and even the General Accounting Office.[Ref 4-37] More importantly, however, MLS technology was no longer the only alternative to ILS.

In 1983, the United States decided to make its satellite navigation network available to the world. The system, which was developed originally by the Department of Defense for military purposes, was based on a constellation of geosynchronous satellites. By using satellitetransmitted radio signals to measure the distance from a receiver to several different satellites, the exact position of the receiver could be determined at all times. The Global Positioning System (GPS) as the U.S. satellite network was called, offered so much more capability and flexibility than any other navigation system that it was described as "the greatest opportunity to enhance aviation system capacity, efficiency and safety since the introduction of radios and radiobased navigation."[Ref 4-38] GPS was not dependent on any ground navigation aids, which meant that it could provide guidance in a much wider variety of locations than navigation systems based on radio transmissions from particular ground positions. GPS could also provide precise position and velocity information in three dimensions, at any instant in time, which other navigation systems could not do.

There were limitations on the system, however. To keep the defenses of the U.S. from being compromised, the Department of Defense intentionally degraded the accuracy of the satellite signals for civil use. As a result, the position accuracy was only within 300 feet, which was acceptable for navigation, but not precise enough for a landing system. These position errors could be corrected and the inherent accuracy of the satellite signals could be improved further through the use of "differential" techniques, although this approach did require ground installations. In a differential system, a stationary GPS receiver was placed at a surveyed location near the airport. The ground site would receive the satellite transmissions, compute how much correction was needed to match its surveyed location, and transmit that correction factor to approaching airplanes through a realtime data link. There were still questions about the capability of GPS to provide enough accuracy for the most demanding instrument approaches, however.[Ref 4-39]

Representatives of the airlines' Air Transport Association (ATA) began to advocate the use of GPS, perhaps in combination with the current ILS system, as a possible alternative to MLS. Some simulation research had also started to evaluate the use of GPS for landing guidance systems, but there was still tremendous momentum within the FAA and the international community for the microwave landing system. The GPS movement got a tremendous boost, however, from a series of flight tests conducted in the fall of 1990 by NASA's TCV/ATOPS program and the TSRV 737 airplane. [Ref 4-40]

Ironically, the TSRV research was not even directed toward civil aviation use of GPS. Honeywell, Inc. had developed a GPS receiver integrated with an inertial reference unit (IRU) that it wanted to test to see if it was accurate enough to be used as a landing aid for returning space vehicles, such as the shuttle or an emergency crew rescue craft. Space vehicles generally land in good weather, on very long runways, however, so the accuracy requirements are not quite as stringent as those for commercial airline operations in extremely low visibility conditions. [Ref 4-41]

Because the ATOPS 737 had extensive experimentation and data collection capabilities, Honeywell proposed testing its GPS/IRU system on the NASA airplane. One of the NASA managers suggested integrating the GPS/IRU equipment with the airplane's existing autoland system, to gather some additional data on the performance of the equipment all the way through landing. The autoland portion of the experiment was almost dismissed, however, because from a research standpoint, there was little to be gained. The Langley engineers knew that the autoland system on the airplane worked, so performing autolands with Honeywell's GPS/IRU equipment appeared to be simply demonstrating a capability that was already obvious.[Ref 4-42] What the research engineers at Langley failed to appreciate was that while that capability was obvious to them, it was not yet obvious to the rest of the world.

In October and the early part of November 1990, the 737 performed a total of 25 hours of flight tests with the GPS/IRU system, using a differential GPS station at the Wallops Island flight test facility. Initial tests with just the GPS/IRU guiding the autoland system showed an unacceptable vertical error, so a radar altimeter was integrated, as well. With that combination, the airplane performed a total of 36 successful GPS/IRU automatic landings. The airplane's performance did not meet the most stringent precision landing requirements, but it met Category I specifications (within 112.2 feet horizontal and 27.2 feet vertical accuracy) and came close to the 33.8foot horizontal accuracy requirements for a Category II landing.

Aerial view of Wallops Flight Facility, where many of the 737 research flights have been conducted.

To the researchers at Langley, the results were interesting, but not earthshattering. In addition to the 36 successful autolands, there were others that had to be completed by the safety pilots, although that was typical of Category I landings. The accuracy of the GPS/IRU/radar altimeter system also still fell substantially short of the microwave landing system.[Ref 4-43] To the outside world, however, the specific error levels and the combination of equipment that the landings required were secondary. The ATOPS program had just demonstrated, in an actual transport airplane, that a GPSguided autoland was possible. In a few test flights, the ATOPS researchers had changed the face of the GPS debate from a basic question of whether GPS could ever be used for landings to more specific questions about the degrees of accuracy and reliability that could be achieved. [Ref 4-44]

The results achieved in the NASA/Honeywell tests in 1990 were not nearly good enough to support using GPS instead of MLS, but they helped build momentum for further research and development efforts. By 1993, GPS technology had already advanced so far that the ATOPS program began another series of test flights with equipment that, in theory, could achieve accuracies not only adequate for Category II/III landings, but within 1019 centimeters.

The 1993 research effort used an Ashtech, Inc. P12 12channel, carrier phase tracking receiver, with additional processing software developed by Ohio University. The Honeywell equipment tested in 1990 was a much earlier generation GPS receiver that used code tracking reception and only two channels. The carrier phase tracking receivers were more accurate, because instead of just locking on to a code associated with the GPS signal, a carrier phase receiver locked onto the actual GPS frequency wave itself. The Ohio University software was designed to improve the accuracy of the equipment even further by locking on to not only the frequency wave pattern, but the exact frequency cycle associated with the airplane's position. [Ref 4-45]

Even if the Ashtech/Ohio University system performed flawlessly, some questions about GPS would remain, especially within the ICAO community. To be acceptable as a landing system, GPS would not only have to have high reliability, it would have to be able to alert pilots instantaneously if a satellite signal was flawed. There were also uncertainties about how easy it would be to jam a GPS signal, and concerns among the international community about relying on a satellite system owned and operated by the U.S. military. [Ref 4-46]

In 1988, the Soviet Union also agreed to make its GLONASS satellite signals available for international civil use, which reduced some of the concerns about satellite availability and U.S. control of the system. ICAO designated an international satellite system, including GPS, GLONASS and any other country satellites that might be added to it in the future, the "Global Navigation Satellite System," or GNSS.[Ref 4-47] The political upheaval in what had been the Soviet Union following 1991 left some uncertainty as to when or if the full complement of GLONASS satellites would be operational, however.

ICAO finally decided to wait until 1995 before determining whether or not to alter its plan to implement MLS. The original deadlines called for installation at all international airports by 1998 and transition to MLS as the primary navigation system by 2000. By 1995, the FAA and NASA were expected to have substantially more information on the potential and limitations of GPS, and ICAO members could determine whether they wanted to extend the deadline for MLS, shift the main navigation and landing system focus to GNSS, or look at some combination of the two.[Ref 4-48]

The landing system decision remained a highly charged issue, however, and there were split opinions both within ICAO and within the United States itself. The United Kingdom, for example, decided to proceed to MLS implementation without waiting for additional research into GPS. The poor visibility conditions in the U.K. made the ability to perform Category III landings a necessity, and ILS frequency congestion and interference problems were more severe in Europe than they were in the United States. The British reasoned that even if GPS technology could be developed far enough to provide the accuracy for Category III operations and the questions regarding system integrity, reliability and availability could be resolved satisfactorily, development and implementation of the system would take too long. [Ref 4-49]

The United States looked much more favorably upon a GPS or GNSS system, but there were still strongly divided views within the FAA, NASA and the aviation community about what the next landing guidance system should be. Most people agreed, however, that even if the MLS system was implemented, it would probably be in a much more limited capacity than originally envisioned. [Ref 4-50]

The 25 year controversy over a new landing guidance system underscored just how much political, financial and other external factors can influence the acceptance and application of new technology. In the mid1970s, the microwave landing system clearly demonstrated its superiority over the existing instrument landing guidance equipment. Yet almost 20 years later, the old ILS system was still in use, and MLS was being overshadowed by an even more advanced technology.

The landing system debate also showed once again the impact an airplane like the TSRV 737 could have on the acceptance of new technology. With the use of the NASA 737, the backers of both the TRSB microwave landing system and the global positioning system were able to go beyond simply describing or simulating the capabilities of the technology. They could actually demonstrate its performance, in an actual air transport class airplane and in realistic flight situations. In the case of the MLS competition, this gave the U.S. candidate an edge the British entry could not match. In the case of GPS, it forced a reevaluation of the system's possible applications and generated momentum for further research and development efforts. The NASA 737 was continuing to prove that actual flight demonstrations could give new technology a level of credibility that no amount of laboratory testing or simulation could achieve.