Perhaps more than any other case in this series of studies, the cellular telephone represents a largely incremental step in a technology development that had been underway since the 1920s. Although several key developments during the 1960s and 1970s led to important supporting technologies that helped make cellular telephone systems a reality; no single technological "breakthrough" is responsible for the appearance of cellular systems, although there were some "ah-ha's" on the business end. As one observer notes, "Cellular radio is not so much a new technology as a new idea for organizing existing technology on a larger scale" (Calhoun, 1988: 39).

Nor was the cellular technology the outgrowth of fundamental research on radio frequency propagation and control. Rather, it was the result of demand-driven technological improvements delivered by corporate researchers, primarily at Bell Labs and Motorola, whose general direction and goals were more or less clear from the beginning. The development of commercial cellular systems did not occur rapidly -- almost 36 years passed between the initial elucidation of the cellular concept at Bell Labs in 1947 and the debut of the first commercial systems in Chicago and Washington/Baltimore in 1983. Much of this delay was the result of regulatory and business decisions that pushed the development of cellular technologies to the back burner. However, the delay ultimately permitted developers to incorporate many new supporting technologies such as microprocessors and integrated circuits, which were developed during this interim period, into the design of the cellular telephone as we know it.

Tracking the generation and flows of knowledge in the cellular phone case came largely through interviews, since the number of publications related to cellular development is relatively sparse. In fact, if one subtracts two major compilations by Bell Labs personnel and two submissions to the FCC, the public literature is somewhat limited. There were very few academic publications during the period covered by the case that deal with cellular or even mobile radio. There is some gray literature in the form of internal Bell Labs and Motorola memoranda and other publications (which we have examined), but little else.

This case ends with the commercial take-off of the cellular phone in the mid-1980s. However, the analog cellular systems that began commercial operation in the 1980s are by no means the end of history. Commercial and technological advances in mobile telephony continue to this day, driven by the seemingly relentless demand for more and more capacity. We provide an epilogue to this case covering selected further developments in mobile telephony, primarily digitalization, spread spectrum techniques, Personal Communications Services (PCS) and other methods for increasing capacity, in order to shed some light on recent developments.
 

II. Defining the Cellular Telephone

From a user's perspective, cellular telephone systems are differentiated from their predecessors (usually known as IMTS or Improved Mobile Telephone Service) by the ability to provide mobile telephone service to a large number of mobile users (particularly in urban areas), as well as to permit these users to move freely and rapidly without significant service degradation almost without geographical limit. The quality of cellular telephone connections exceeds that of predecessor systems and frequently approaches that of land line telephone conversations. Cellular systems also facilitate the widespread use of portable mobile phones, as opposed to larger vehicle-mounted units that were the most common type of mobile unit in the pre-cellular era.

Cellular systems are able to accommodate a large number of users within a given geographic area even while using a rather limited portion of the frequency spectrum. This high capacity is achieved by using a large number of relatively low-powered base station transceivers to provide service to a small geographic area known as a "cell." Because the transmitters are low-powered, and because FM provides a "capture effect" to suppress interference, their frequencies can be reused in nearby (although not adjacent) cells, thereby multiplying the practical number of channels available in a large service area. Furthermore, as demand increases cells can be subdivided or split into smaller units with even lower powered transmitters, further increasing the number of times given frequencies can be reused within a given service area. As users move about the service area, they are switched from one base station to another through a complex and sophisticated system known as "handoff." Handoffs are accomplished in a nearly instantaneous manner so that users are unaware that their mobile telephones changed frequencies and have been switched to another base station. A computerized control system monitors and directs the entire network, identifying mobile units within the service area and establishing connections over which conversations can take place.

At this point, it may be useful to keep in mind that cellular systems generally involve two types of radio transceivers -- base stations and mobile units-- and that mobile units are further subdivided into larger units mounted in vehicles, and portable units meant to be carried anywhere (including into vehicles) by their users. Cellular's predecessor systems as well as some early cellular systems operated with the concept of providing service almost exclusively to vehicle-mounted units. However, the use of portable phones has grown very rapidly so that they now comprise the majority of cellular mobile units being sold. Portable units have different, more demanding, technological requirements than vehicle mounted units, which can be relatively large and heavy, and which do not have significant limits on power consumption. Since portable units have become the most common type of cellular mobile unit, we are including some technologies essential for them, but not required for the earlier vehicular units, among the technologies we discuss here.

From a technological perspective, many modern innovations can be characterized by a number of intrinsic and supportive technologies that are combined to make them possible. However, some observers believe that there are no new technologies intrinsic to the cellular telephone, which they characterize as just the use of existing technologies in a smarter way. Other observers have identified a number of technologies as intrinsic to cellular telephony, but there is no strong consensus on which ones they are. Therefore, classification of cellular telephone technologies into intrinsic and supportive categories, an analytical device that proved helpful in earlier cases, is somewhat more arbitrary here. The technologies that we have labeled intrinsic to current cellular systems are (in approximate order of importance): 

• The cellular concept itself, with a system architecture that includes:
• Frequency reuse within a relatively small geographic area using low power transmitters
• Automated central system control including "handoff" as mobiles move from cell to cell (the key difference between cellular and other systems)
• Cellular geometry consisting of hexagonal shaped coverage zones (or cells) and design software to simulate RF propagation and facilitate system layout
• Cell splitting, i.e., putting new cell sites between previous ones, to increase capacity as the system becomes loaded
• The concept of channel trunking combined with low-cost frequency synthesizers (developed for predecessor mobile systems, but essential for cellular) as well as low cost crystals with new levels of precision for generating reference frequencies

The history of the cellular telephone begins with the history of mobile radio in the 1920s. The first land mobile systems were used by public safety agencies, primarily police departments. The earliest system was tested in Detroit beginning in 1921. It consisted of police calls interspersed in a regular commercial program ("Calling all cars...") that broadcast instructions to police in vehicles. If the police officers wanted to talk back, they had to stop at a police telephone and call in. During the 1930s, two-way systems came into use as transmitters were built that could be operated in vehicles, and the advent of frequency modulation (FM) provided much clearer conversations free from vehicular static. Almost all mobile systems operated below 40 MHz, since little was known about propagation at frequencies above that range, particularly in urban environments. However, research on the propagation of higher frequencies continued almost continuously from that time on.

World War II demonstrated the superiority of FM transmission -- only U.S. forces used significant numbers of FM battlefield systems -- which proved easier to use and more difficult to jam than AM systems. Following WW II, many servicemen returned to civilian life with a knowledge of radio technology and an appreciation of the value and convenience of mobile radio communications. Surplus military radio equipment, particularly the Motorola "Handie-Talkie," entered civilian life as taxi dispatch radios, particularly in New York.

The late 1940s were important years for mobile radio. AT&T introduced the first commercial land mobile radio telephone system in St. Louis in 1946. However, the service was limited by a lack of communications channels (frequencies), and the systems were cumbersome to use, with "push to talk" features and manual connections via operator. Nonetheless, 25 U.S. cities had mobile service by year's end.

Research on mobile communications at Bell Laboratories, which had been underway since the 1930s, continued. In 1947 the cellular concept "materialized from nowhere" and was embodied in an internal Bell Labs memorandum authored by D. H. Ring (1947) with major input from Bell Labs colleague W.R. Young. This paper summed up the thinking of Bell scientists, and suggested that it might be possible to build a high capacity land mobile radio telephone system that could provide wide area coverage with a modest allocation of frequencies. It identified many of the concepts essential to modern cellular systems, including the critical idea of using low power transmitters for small areas (the term "cell" did not come into common use until almost 20 years later) permitting significant frequency reuse within the service area. It identified the hexagonal layout characteristic of an idealized cellular system layout; identified the concept of dividing areas into even smaller areas (cell splitting) to increase capacity when demand increased; and identified the need for "handoff" techniques. Finally, it noted that base stations could control the transmitter power of the mobile transmitters to reduce interference. The Ring memo was not a system design, but a suggestion of additional areas of research required to verify whether the concept could, in fact, be implemented.

That same year, AT&T developed and proposed to the FCC the idea of "trunking," by which a number of radio channels are grouped together so that any mobile unit wishing to communicate could search the trunk group for an available channel. The practice of trunking was important for existing land mobile radio systems and would become indispensable for cellular systems. AT&T also petitioned the FCC for additional frequencies for mobile radio service, but in a key event impacting the development of the cellular telephone, the FCC declined to act, preferring to reserve the requested spectrum for the burgeoning television industry. Deprived of the necessary frequency allocation in which to implement potential dramatic improvements to mobile radio, Bell Labs put research on cellular systems on the back burner. Some research on propagation and further development of the cellular concept continued, but because of the FCC's action, there was little impetus for rapid development. Bell's mobile radio research focused on further development of its existing mobile radio products as well as propagation studies, and studies on the concept of "diversity," or the use of two radio paths (i.e., using two antennas, two receivers, etc.) to guard against outages caused by poor signal conditions.

The decade of the 1950s saw few new developments in cellular. The FCC made a minor addition to the frequencies allocated to mobile radio service, but again turned down another AT&T request for a significant band (75 MHz) at 800 MHz to be used for mobile radio. Improvements in the stability of oscillators allowed FM channel spacing to be reduced to 60 KHz. However, receivers in mobile radio systems of the day were not able to discriminate between channels separated by only 60 KHz, requiring that the adjacent channels be assigned in alternate cities. In private dispatch systems, continuing advances permitted separations as low as 25 KHz in the 450 MHz band.

At Bell Labs, scientists and engineers continued to think about how one might implement high capacity land mobile radio telephone systems based on their earlier conclusions embodied in the 1947 Ring paper. In internal Bell Laboratory technical memoranda written in 1958 and 1959, the desired characteristics of and problems associated with a future system were being discussed, as well as some of the proposed technical solutions. However, effective implementation could not come until the 1970s, when the appearance of microprocessors and economical computers became generally available, and specific programs for managing handoff had been written for them.

In the 1960s, the pace of development of land mobile radio increased as key technical and regulatory developments set a course for the development of cellular systems. A reduction in FM frequency deviation from 15 KHz to 5 KHz plus a general tightening of filter specifications progressed to the point where channel spacing could be reduced to 30 KHz, and adjacent channels could be used in the same city. (O'Neill, 1985: 412) These reductions in FM channel width made the use of the FM spectrum efficient enough for widespread commercial use. Without the efficiencies brought about by these channel size reductions, widespread cellular service would have been difficult. (Calhoun, 1988: 32)

Work by the Bell System and Motorola, the major supplier of mobile equipment, continued to advance the capabilities of mobile telephone equipment and systems and led to the Improved Mobile Telephone Service (IMTS). IMTS relied on a number of recent technological advances, particularly the reductions in FM channel spacing noted above, as well as low cost, low power, small size and more complex circuits made possible by new semiconductors. Automatic channel selection made use of the "marked idle" concept, in which one of the available channels was marked with a tone. for which the mobile station would search automatically. IMTS permitted customers to dial calls, and eliminated the need to "push to talk," while retaining backward compatibility with older systems. Widespread commercial introduction began in 1964.

Demand for mobile telephones continued to exceed the abilities of IMTS. In the mid 60's, Bell Labs, looking for additional technological means to expand mobile capacity, dusted off the cellular concept and assigned a team to provide an assessment and plan for development. However, as this effort was getting under way, the U.S. Department of Transportation approached AT&T to develop an on-board telephone system for the new Amtrak Metroliner. The Metroliner project lasted from mid 1967 to early 1969 and produced a primitive forerunner of later cellular systems, although it did not employ multiple channels, frequency switching and complex handoff systems. Only a small number of channels were used, spaced geographically up and down the rail line, with periodic reuse of frequencies. As the train moved, sensors along the track caused landside units to switch on and off with the movement of the train, effectively passing control of the train unit from one landside transceiver to the next without changes in frequency. Coverage was even provided in tunnels by transceivers located at either end. The Metroliner mobile telephone service remained in operation until 1983.

Frequency propagation studies continued in the 1960s and in 1968, a landmark study, "Field Strength and its Variability in VHF and UHF Land Mobile Service," was published by Y. Okumura. Based on an extensive project to conduct field strength measurements in Japanese urban settings, the study was carried out in preparation for the implementation of an NTT cellular system (originally planned for 1968, not actually implemented for another decade). The paper, supplemented by studies carried out in the U.S. by Bell Labs, Motorola and others, became the basis for the design of several computer-modeling systems developed to predict frequency propagation characteristics in urban areas where cellular systems were being implemented. These computer systems (the two main cellular players, Bell Labs and Motorola each developed its own) became indispensable to the design of commercial cellular systems. In addition, it was becoming evident from research at Bell Labs and elsewhere that channel (frequency) synthesizers could be built to generate any of hundreds of frequencies; that antenna combiners for many channels at a base station could be designed; that stored-program electronic switching could make handing off a call from one base to another practical (a key development); and that signaling could be fast and reliable. (O'Neill, 1985: 415)

Throughout the 1950s and 60s, the FCC had resisted continuing requests for additional frequency allocations for land mobile service. However, in 1964, the Commission did convene the "Advisory Committee for Land Mobile Radio Services" to examine the congestion in land mobile telephony. The FCC was seeking a technological solution to the need for additional bandwidth, and the Committee ultimately recommended that a substantial portion of the underused UHF TV band (the 800 MHz band) be reallocated to public and private land mobile communications (i.e., cellular and dispatch systems). In 1967 the FCC announced it would study the feasibility of these recommendations and subsequently opened docket 18262 (known as the "cellular docket"), which proposed a frequency reallocation of 75 MHz (taken out of TV's UHF band) for mobile service.

The FCC's First Report and Order in docket 18262 reallocated 115 MHz in the upper portion of the TV UHF band and set aside 64 MHz for "land mobile communication" (i.e., cellular). In addition, 40 MHz were allocated for private mobile services (primarily dispatch), and 11 MHz for air to ground service (which never materialized). The FCC asked for recommendations on how this spectrum could be used most efficiently. The FCC's willingness for the first time to allocate a significant amount of new bandwidth for mobile radio service moved development efforts back to the front burner. In 1970, AT&T was authorized to develop and test the cellular concept within a testbed system in real urban conditions in Newark and Philadelphia.

In response to docket 18262, AT&T submitted a technical proposal to the FCC in late 1971. The report covered not only the technology of a cellular system, but service features, coverage, capacity growth, customer opinions on quality, and costs as well. (O'Neill, 1985: 415) This proposal (based on an earlier Bell Labs report to its AT&T parent) was one of the key documents in the development of the cellular telephone, and outlined AT&T's technical and commercial vision of a cellular mobile system, referred to as Advanced Mobile Phone Service or AMPS. AT&T reported in the proposal that the field tests of the prototype systems in Newark and Philadelphia were successful, with cell splitting into smaller cells having radii as small as 1.4 miles. Docket 18262 and the AT&T proposal had serious implications for who would dominate the future of mobile radio service. In response, Motorola, at that time the major player in mobile radio equipment and dispatch systems, launched its own research program to develop plans for a cellular system. These were presented to the FCC 18 months later (in April 1973) in a document entitled "The DYNA T.A.C Concept."

Although both the AT&T and Motorola visions of a cellular mobile radio telephone system embodied most of the key concepts of today's cellular service, they differed in a number of assumptions and approaches to solving the technical problems of cellular service. These differences ultimately required resolution and technical compromise before a cellular standard was adopted. AT&T's concept for a cellular system focused initially on service primarily to vehicular mobile units, with mobile units to be added later. In the early cellular era, AT&T lacked the experience with portable units which was Motorola's strength, and also relied on a vehicular electrical system to provide the power needed to operate transmitters in a large area system. On the other hand, Motorola, long a manufacturer of portable units, was said to have a "portable culture," and proposed a system which included a large number of portable hand-held units from the very beginning.

In addition, Motorola was convinced that frequency deviation could be held to a lower level than AT&T believed, and therefore channel spacing could be narrower than AT&T had proposed. AT&T had proposed a channel width of 40 KHz, while Motorola proposed 25KHz. (30 KHz was ultimately adopted.) AT&T proposed a frequency deviation of 12KHz while Motorola proposed 5 KHz. (8 KHz was ultimately adopted.) Both AT&T and Motorola ultimately tested their concepts for cellular service in real urban conditions (AT&T in Chicago and Motorola in Washington/Baltimore). The differences between the technical standards of the two systems were later resolved at meetings held under the auspices of the Electronics Industries Association from which an industry standard emerged and which the FCC ultimately adopted.

By the early years of the 1970s, most of the major technical hurdles in the development of the cellular phone had been surmounted. Throughout the 1970s, the cellular telephone activity was on the regulatory and business front. Following the successful AT&T tests, FCC made an allocation of 40Mhz of spectrum to be given to a single wireline common carrier (WLCC) for cellular service in each market. In addition, 30 MHz was allocated for private mobile systems (primarily dispatch systems) and 45 MHz was reserved for future growth. The FCC specifically stated that it wanted cellular systems to be compatible with each other so that customers could "roam" from one city to another using the same car phone. The decision to limit cellular licenses to WLCCs (i.e., the Bell System and the smaller telephone companies, such as GTE) sparked considerable protest among the Radio Common Carriers (RCCs). The RCCs consisted of a large number of mostly small firms engaged in radio-related businesses such as paging. They viewed the cellular business as an opportunity that should be open to them as well, and following well-established tradition, they filed suit.

In partial response, the FCC revised its 1974 decision to permit any qualified carrier to enter cellular service. The RCCs sued again, charging that although the 40 MHz allocation was available to any qualified carrier, the decision would still result in a Bell monopoly since they owned all the required technology. In spite of this assertion, in 1975 one of the RCCs, American Radio Telephone Service Inc., (ARTS) with Motorola as a partner, filed an application with the FCC for a developmental cellular system in Washington/Baltimore. At about the same time, AT&T, through Illinois Bell, applied for a similar developmental system in Chicago. Both of these applications were granted in 1977, and two years later, after Congressional hearings and other legal and political maneuvers, the FCC finally decided to consider granting two licenses per market, one for the local WLCC, and one for any other qualified common carrier.

In 1979, the first commercial cellular system, owned by NTT, the Japanese national telephone company, began operation in Tokyo. Cellular systems in the Nordic countries also entered service about the same time. The development of cellular technology outside the U.S. did not encounter the same level of regulatory barriers that delayed the commercialization of cellular within the U.S. In addition, in the Nordic countries, an aggressive cooperative research program consisting of local academics (who were much more involved in mobile radio than their U.S. counterparts), Eriksson and Nokia drove the development of the Nordic system.

Also in the early 1980s, the first commercial digital telephone switches began to appear. The first digital telephone switch installed on a local system was marketed by Northern Telecom, although AT&T already had a digital toll (long distance) switch. The advent of commercially-available digital switches deprived AT&T of the technological monopoly in analog telephone switching systems it had held up to that time. As a result, an RCC or any other carrier could purchase a digital switch, effectively eliminating one of AT&T's major technological advantages. On the regulatory front, the FCC began another round of cellular rule making in Docket 79-318 to establish general policy and rules for commercial cellular service. At that time, more than 50,000 people were on the official waiting list for IMTS (Motorola Cellular Telecommunications Milestones).

In 1982, the FCC released its final Report and Order for Docket 79-318, and adopted rules creating a commercial cellular radio telephone service. These specified that there would be for two competing cellular systems in each of 90 urban markets, one owned by a nonwireline company (known as the "A" side), and one owned by the wireline company/local phone company (known as the "B" side). The FCC decided to conduct detailed comparative hearings to select the most qualified applicant, where more than one company applied for the same license. Each licensee was to get 20 MHz of the 40 MHz allocated for service in each area. This allocation would accommodate 333 full duplex channels, each 30 KHz wide. The FCC began accepting applications for the 30 largest markets, and received more than 200 applications, many from major firms with significant communications experience.

The FCC administrative apparatus was overwhelmed by the sheer number of applications. By October 1983, almost 700 applications for cellular service had been filed for the bottom 30 markets (ranked 61st to 90th in size). The FCC amended its rules in May, 1984 to specify that lotteries would be used in all but the top 30 markets to select among competing applicants, since the number of applicants was so large, and hearings would take too long. A number of speculators with no interest in operating cellular systems won lottery allocations and became wealthy selling their rights to others. (Grant, 1994: 352)

On October 13, 1983, the first pilot cellular system began operating in Chicago. The second system, in Baltimore/Washington, was activated on December 16, 1983. These pilot system tests differed from earlier tests in that they were larger in area, used commercial equipment, and involved real, fee-paying users. The pilot systems was so successful that some customers refused to return their pilot test phones unless they were guaranteed access to commercial service when it began.

On January 1, 1984, just as cellular had begun to arrive in the marketplace, the settlement of the antitrust suit between the Bell System and the Department of Justice was carried out. The execution of the settlement broke up the Bell System into a number of local parts and an 'all other' part (including long distance). The Bell telephone companies were rearranged into seven regional units (Regional Bell Operating Companies or RBOCs). Cellular service wound up with the RBOCs, whose geographical boundaries made no sense for cellular. For example, to expand the New York City system to cover northern New Jersey required the approval of the court in proceedings that often required months or years. The Bell regional companies quickly began acquiring nonwireline cellular companies outside their regions (a practice allowed by the Bell System breakup agreement). Five of the Bell regional companies soon acquired nearly one-third of the nonwireline cellular companies outside of their respective regions in competition with their former "siblings."

In 1984, Motorola shipped the first commercial portable cellular telephone (with a suggested price of $3,000 - $4,000). Within the year, Motorola had introduced a line of cellular products including portable telephones, mobile (vehicular) telephones, as well as transportable telephones and was producing 43,000 cellular telephones annually. (Motorola Home Page Motorola Milestones) By the end of 1984, Motorola's sales of cellular phones reached $180 million annually from essentially zero the year before. Also by the end of 1984, Washington/Baltimore had two competing cellular providers, the original ARTS system, and the wireline service operated by Bell Atlantic. By the following year, 82 cities had two-provider cellular service.

As early as 1988, some cellular systems (particularly New York and Los Angeles) were already becoming overloaded as the promise of nearly infinite expansion of capacity from cell splitting turned out to be more costly and difficult than foreseen. Nonetheless, cellular expansion continued unabated, and by 1990, cellular construction permits had been issued for at least one system in every market in the United States, and the cellular subscriber count topped 5 million. Two years later the number of cellular subscribers had doubled, by 1995 had reached 25 million, and it is currently estimated to be just over 50 million.
 

Intrinsic Technologies

As noted above, defining the intrinsic technologies of the cellular telephone has been somewhat more difficult than for the other cases. Nearly all the sources interviewed for this case, as well as the literature (see the Calhoun quote in the Introduction), characterize the development of the cellular phone as an evolution of existing technologies without any breakthrough discoveries. Some believe that the cellular telephone contains no intrinsic technologies at all. This view maintains that all the technologies incorporated into the cellphone were part of the continuing evolution of mobile radio telephones, and that many of the essential technologies in today's cellphones (e.g., microprocessors, integrated circuits) were developed for other reasons and utilized by cellphone developers as they were utilized in so many other areas of electronics. Nonetheless, we have found it useful to identify a number of technologies that set the cellular telephone apart from predecessor mobile systems. There is modest support for the assertion that the intrinsic cellular technologies are:

1. The cellular concept itself, including:

Frequency reuse in numerous small cells served by relatively low power transmitters. Even this key element of cellular system design is not entirely new. The principle of frequency reuse underlies the system by which broadcast frequencies have been assigned around the nation since the beginnings of the broadcast media. It recognizes that a more efficient use of the available frequencies will be achieved by reassigning them in geographically separated markets, thereby reusing each channel many times. These frequencies are reused, not in adjacent markets where interference might occur at the boundaries, but in markets separated by one or two others. Thus Washington's TV channel 7 can be safely reused in Philadelphia, but not in Baltimore, which is close to Washington, and which would cause users to be plagued with interference.

As noted, earlier, Bell Laboratories scientists and engineers realized as early as 1947 that this same frequency reuse principle could be applied to mobile telephone service in specific (usually urban) markets (although the reasons for signal dissipation differ from TV to mobile radio). Mobile radiotelephone services of the time, as well as private services such as dispatch systems, relied on a single high power transmitter for a large service area with a number of relatively high powered (by cellular standards) mobile units communicating with it over a set of one or more frequencies.

The Bell radio engineers theorized that if a set of channels were used with a low power transmitter, they could be reused many times, although not in adjacent cells, where they would interfere at the cell boundaries, but rather several cells away. The diagram demonstrates the principle. Frequencies in cell 1 in the diagram are reused in other non-adjacent cells (also marked 1). The distance between the cells using the same frequency set is determined by the ratio (D/R) of the distance to the nearest interfering cell, or D, to the distance from the transmitter (i.e., the cell radius), or R, which is 4.6 in the diagram. Determining the optimum value of D/R is important, since lower values are functionally equivalent to having additional frequencies in the system. In an urban area, the power received at a mobile unit is proportional to the distance to the transmitter to the fourth power (R⁴). It can be easily shown that the signal to interference ratio (S/I) equals D/R to the fourth power, and it can also be shown that D/R for a perfect grid of base stations is proportional to N, the number of sets into which the allocation is divided. The value of D/R can be adjusted (since it is a function of the power of base station transmitters that determine the cell radius) in order to achieve the desired signal to interference ratio (S/I). However, there is little leeway in designing cellular systems as the quality degrades very quickly as smaller values for S/I are introduced into the design.

A number of techniques have been introduced to further increase the capacity of cellular systems by increasing the number of times that frequencies can be reused in a system including the use of cell sectoring and directional antennas. Early cellular designs employed omnidirectionalantennas. These antennas meant that all the channels in a cell's set would be equally likely to interfere with the same frequencies in nearby cells reusing the same channels. However, the use of directional antennas (which were commercially available at the time) to effectively subdivide the cells into sectors (usually three) enabled subsets of the cell's channels to be directed toward only a third of the serviced cell, thereby increasing the S/I ratio by a factor of 3. Coordination of the directions of antennas throughout the system further reduced the chances of interference.

Handoff and central control If there is any technology that separates cellular from all the mobile radio systems that preceded it, it is the concept and technology of handoff. As mobile units move in a cellular market, they will, from time to time, cross the boundaries between cells (particularly in areas with small cells). As this occurs, they must be switched to a new channel in the new cell. This "handoff," process is simple in concept, but complex in practice. Handoff occurs when the cellular system's central controller determines that a mobile unit is experiencing low signal strength on its assigned channel and sends a signal to the mobile unit to change its transmit and receive frequencies into the channel set of the new cell. In early cellular designs, it was assumed that it would be necessary to monitor the physical location of each mobile unit and assign it to the geographically nearest cell base station to prevent mobile units from interfering with other units in nearby cells. However, after the early trials, it became evident that signal strength alone, not physical location, was the best criterion for making handoff decisions.

Handoff is coordinated by the central controller's computer, and is generally based on frequently sampling the strength of each mobile unit's signal at nearby base stations. The central computer causes handoff to occur by sending a signal to the mobile unit to switch to an available channel in the new cell. Since cellular systems can have hundreds of channels, frequency synthesizers (see below) which could accurately and quickly generate the new channel were essential to cellular systems. The actual handoff takes place very quickly and is usually unnoticed by the user. In early experimental systems, handoff occurred whenever a mobile unit encountered a base station stronger than the one it was currently communicating with. However, designers soon discovered that mobiles traveling near cell boundaries were being handed back and forth several times a minute as signal strength went up and down. Subsequent system designs maintained the connection between a mobile unit and the base station until the signal strength began to enter the unacceptable range, at which point handoff would occur.

Cellular geometry and system design is another critical element of cellular systems. The placement of adjacent base stations determines the shape of the cells used in a system. The 1947 Bell Labs paper discussed the three possible geometric shapes, the triangle, the square, and the hexagon. All three shapes have the same distance between the center and the farthest point (the farthest a mobile unit could be and still occupy the cell and thus the point of weakest signal). However, of the three shapes, the hexagon is slightly more efficient, permitting the largest number of mobile units to be reached from a single base station (MacDonald, BSTJ 1/79, p.20). Of course, real life cells are not perfect hexagons - they look more like fat amoebae. The figure shows a "real" cellular layout.

Designers of cellular systems need a way of optimizing the placement and positioning of base stations, since deviations from theoretical hexagonal geometry must always be made to accommodate practical antenna placement requirements. Furthermore, terrain and other factors can result in weak signals in certain areas, a situation that needs to be remedied by changing the placement of base stations. Early cellular designers knew that they would not be able to efficiently design a commercial cellular system without some sort of automated tool. In fact, costs for "hand designed" cellular systems would have been very high, severely hindering the general introduction of cellular technology. Therefore, both the two main players, AT&T and Motorola, designed cellular simulation software as an essential part of their work on cellular phones. This software was largely based on the results of propagation studies, primarily the 1968 Okumura work, supplemented by their own measurements. As third party players began to enter the cellular game in the late 1980s, commercial cellular simulation software was produced by commercial vendors (often based on work done at universities).

Cell splitting is the process of dividing cells into smaller cells in order to increase capacity as the system grows. In smaller cells, which have even lower base station power (in cells of all sizes, mobile unit power output is adjusted by the system's controller to maximize quality and minimize interference), channels can be reused in the same way as with larger cells, but at a much closer physical distance, sometimes only a few miles. The diagrams show how cells can be subdivided in stages and that the process does not have to be symmetric. Larger cells can continue to exist until their entire territory is a filled, step-by-step, with smaller cells.

While simple in theory, cell splitting turns out to become more difficult in practice. As the physical size of the cells decreases, siting of the base stations becomes more problematic, since there is less room for deviation from the base station's ideal location. There are substantial technical and business problems locating antennas in urban areas such as local objections to the visibility of antenna systems. The reduced leeway for base station locations in small cells makes small cell design more difficult to carry out from both business and technical perspectives. Furthermore, although the technique of cell splitting permits cellular systems to start with limited investments in equipment and then grow slowly as use increases, economies of scale are limited because the cost of base stations is relatively constant regardless of their power output.

2. Trunking and Low Cost Frequency Synthesizers

Trunking was first proposed by AT&T in 1947 and was used for all subsequent mobile telephone systems. Trunking is the practice of grouping a number of channels together so that units in a given radio system can use any available channel in the trunk. When a mobile unit wants to begin a communication, it scans the channels in the trunk for an available frequency, and seizes it for the call. At first this process was manual, then became automatic as idle channels carried a specific tone for which mobile units would search, seizing the first available channel so marked. Cellular systems which can use hundreds of communications channels employ a separate control channel for passing instructions from the system controller to the mobile units telling them what channels in the trunk to use.

Large, multi-channel trunks require very fast frequency synthesizers that can switch mobile units to the initial frequency, and then to new frequencies as the unit is handed off. Low cost frequency synthesizers evolved from large, expensive laboratory units, and first appeared in mobile radio in an improved model of the IMTS mobile radio system. The early IMTS mobile units had manual channel selection buttons and crystals for each frequency. Later models introduced the first small, low cost frequency synthesizers that permitted automatic channel selection. Cellular systems have hundreds of channels and require frequency synthesizers that can alter the mobile unit's frequency to one of a large number of choices very quickly. In addition, low cost quartz crystals with new levels of precision for generation of the reference frequency became another important part of modern cellular units.
 

Supporting Technologies

A large number of other recent technologies are employed in commercial cellular systems. Without them, practical portable cellular phones would not have been possible, and cellular systems would have been more limited than those that finally emerged in the 1980s. The most important supporting technologies are:
 

. Solid state technologies, including microprocessors and integrated circuits, permitting electronic circuit switching as well as low cost, low power, high reliability systems (especially for portable units) were critical to cellular systems. These technologies, whose development trajectory was more or less parallel to cellular development, came at just the right time to permit the implementation of fully digital circuit switching, which is critical to the implementation of workable handoff procedures, one of the keys to cellular telephone systems. The advent of integrated circuits also permitted the construction of small, low power and lower cost equipment, particularly hand held portable units. They permitted a much lower cost structure for cellular systems, as well as virtually every other system into which they were incorporated. They also permitted small hand held units to be designed, manufactured, and sold at prices that eventually fell to the point where these units became the predominant type of mobile unit.

. Improved high capacity batteries for portable units. High capacity batteries were critical to the general use of portable cellular phones. Motorola, which did most of the development work on portable phones, licensed battery technology from Texas Instruments and built a battery business (which was only marginally successful). Good portable batteries did not generally become available until lithium ion technology came along. In addition to batteries, smart chargers were also critically important, since most batteries had "memory" and overcharging problems.

. Digital telephone switching. Switching the telephone calls to the land line system was, of course, critical to any large mobile radiotelephone system. AT&T then had a virtual monopoly on (analog) switching technology. However, the introduction of digital switches by non-Bell System providers (just about the time cellular systems appeared) made it possible for non-Bell companies to enter the cellular business without having to partner with the Bell System to obtain their switching technology.

. The characterization of (especially urban) frequency propagation was extremely important for the development of earlier mobile radio systems and paved the way for cellular. Early mobile radio systems used much lower frequencies than the 800 MHz band now used by most cellular systems. In the 1930s, as a result of urban propagation studies it was learned that the 150 MHz band could be used for urban communications, with manageable interference problems. However, little was known about the propagation of frequencies higher than that. In the 1960s it became well understood that radio, like light, had both a geometric (i.e., shadow) and a physical (wave-related) component. The shadows are caused by buildings and hills in the radio path. The wave nature of radio causes fades every half wavelength (every 6-12 inches at 850 MHz, where cellular systems are located) because so many radio paths are summed with random phase and random amplitude. Continuing work led to the characterization of propagation effects at higher frequency bands. Also in the late 1960s, the Okumura paper characterized urban frequency propagation in the VHF and UHF bands (the latter included the 800 MHz band now used by cellular systems).

IV. The Flows of Knowledge

As we have noted earlier, most of the activity leading to the development of the cellular phone took place in industry, first in Bell Laboratories, and later in Motorola and a few other communications firms. Only a handful of academics were working on mobile radio or other topics related to cellular or its intrinsic technologies. Some of the important supportive technologies, particularly microprocessors and ICs, are a different story, but their development is extensively explored elsewhere and tangential to our task here.

To a limited extent, the development of knowledge used in cellular phone development can be tracked in publications by the developers, particularly those at Bell Labs. Beginning in 1947 until 1970, there were a number of internal Bell Laboratories Technical Memoranda on topics related to cellular, usually referred to as "high capacity" or "multi-area" mobile radio telephone service. In general, these papers did not report major research efforts or test results, but summarized the collective thinking of the small teams working on the problem of developing high capacity mobile radio.

The earliest written description of the cellular concept appeared in a 1947 Bell Labs Technical Memorandum authored by D. H. Ring. The TM detailed the concept of frequency reuse in small cells, which remained one of the key elements of cellular design from then on. The memorandum also dealt with the critical issue of handoff, stating "If more than one primary band is used, means must be provided for switching the car receiver and transmitter to the various bands." Ring does not speculate how this might be accomplished, and, in fact, his focus was on how frequencies might be best conserved in various theoretical system designs.

Because it appeared that sufficient frequencies would not be allocated for mobile radio, the 1950s saw only low level R&D activity related to cellular systems. Nonetheless, this modest activity resulted in additional Technical Memoranda in 1958 and 1959, respectively, "High Capacity Mobile Telephone System - Preliminary Considerations," W.D. Lewis, 2/10/58; and "Multi-Area Mobile Telephone System," W.A. Cornell & H. J. Schulte, 4/30/59. These two memoranda discussed possible models for cellular systems and again recognized the critical nature of handoff. In the 1959 memo, the authors assert that handoff could be accomplished with the technology of the day, but they do not discuss in detail how it might be implemented. A review of the citations and acknowledgments of these two papers indicates that they are all to other Bell Labs authors (writing both in internal technical memoranda and in the Bell System Technical Journal). There is a single exception - the 1959 memorandum cites a 1956 paper from Journal of the Institute of Electrical Engineers by J. R. Brinkley on a method for handling data messages to mobiles in different zones (a method that was never employed).

In 1974, William C. Jakes, one of the radio pioneers at Bell Labs, edited a volume entitled Microwave Mobile Communications, a compendium of current knowledge about mobile telephony. All the chapter authors were Bell Labs scientists. The volume, considered a classic, was published as Bell Labs was in the middle of its continuing tests on the experimental cellular test bed system in Newark, and three years after it presented its vision of a cellular system to the FCC in 1971. Most of the book deals with high frequency radio issues such as multipath interference, signal variations, antennas, noise and interference, and diversity systems and techniques. Only the last chapter, "Layout and Control of High-Capacity Systems," focuses on issues, such as system layouts, which are exclusively germane to cellular systems.

A statement by Jakes in the volume's preface is illustrative of the origins of knowledge about mobile radio and cellular: "The authors have drawn freely on published literature that is available and germane to the subject. The bulk of the material presented is based on work done by our colleagues in the Bell Telephone Laboratories, where research directed specifically toward high-capacity systems has been underway for the past decade." (Emphasis added)

In January 1979, Bell devoted an entire issue of the Bell System Technical Journal, (BSTJ) to the cellular telephone, describing the results of Bell Labs' efforts to date to conceive, design, and install cellular systems. Examination of the references cited in the chapters of this issue indicates that most of the sources of knowledge cited in them were from other Bell Labs staff. The references (all to other Bell authors) in the chapter describing control architecture, in which handoff is included, indicate that the details of the handoff procedure were published for the first time in 1971. ("Supervision and Control Features of a Small Zone Radiotelephone System," IEEE Trans. Veh. Tech., 20, No. 3 (1971), p.75-79).

During the course of this study, we interviewed two academics who were working in mobile radio or related areas as early as the 1960s. They were consistent in their view that there were only a very few other academics working with them in that area. One of our interviewees said being an academic in mobile radio was like being "lost in the desert," noting that it was nearly impossible to find other colleagues with whom to discuss one's work in mobile radio. Another source described the area as a "grubby" backwater, implying, correctly we believe, that the subject matter had little attraction for most electrical engineering departments. One of our interviewees said that the situation was somewhat similar at Bell Labs -- until the cellular takeoff period, cellular researchers constituted a tiny fraction of Bell Labs' thousands of scientists and engineers, and worked in what was considered by many to be a professional backwater.

The number of academic publications related to cellular in the 1960s and 1970s is also small. Interviews with Bell scientists and engineers indicate that they were aware of some of this work, and even had occasional contact with these academics. However, this academic work generally had thrusts and goals different from those of AT&T, and thus was only of passing interest. The overall number of publications related to mobile radio and cellular from any source was further limited by the fact that there were few outlets in which mobile radio researchers could publish. At that time, the IEEE Vehicular Technology Conference was heavily skewed toward current systems, and had only a single session on mobile radio telephony. Since the conference would only accept a limited number of Bell Labs papers (presumably to maintain balance), the flow of technical papers from Bell Labs scientists and engineers was somewhat more limited than what it might otherwise have been. Competitive factors also limited the number of papers submitted by Bell Labs scientists and engineers.

One key publication during this period was the landmark study, "Field Strength and its Variability in VHF and UHF Land Mobile Service," by Y. Okumura, published in September 1968 (Okumura, 1968). It was based on an extensive project to conduct field strength measurements in Japanese urban settings sponsored by NTT in Japan as part of a project to implement a cellular system in Tokyo as early as 1968, but in fact the system was not installed for another decade. As noted above, the data from this study made a significant contribution to the cellular system simulation software developed by Bell Labs and Motorola. Okumura visited Bell Labs and Motorola during the early 1970s.

In 1972, 1974, and 1976, three Mobile Communications Symposia were held in Boulder, Colorado under the sponsorship of the IEEE Vehicular Technology Group and the National Bureau of Standards. These symposia provided a new forum for the output of mobile radio researchers. Symposium proceedings were not published (only abstracts were circulated to attendees) in order to be able to highlight the most current work. Most of the presentations focused on the radio link aspects of cellular, according to one of our interviewees who attended, and, as might be expected, proprietary material was not presented. Eventually, when cellular became a "hot" subject, the Vehicular Technology Conferences began to provide expanded opportunities for mobile radio and cellular experts to present and publish, and the Boulder symposia ended. Presentations at the Boulder symposia were primarily by scientists and engineers from industry, and primarily from Bell Labs (roughly half of the presentations). Scientists and engineers from Motorola (constituting roughly 20 percent of the attendees), a few other companies involved in mobile communications, a few scientists from government, and a small number of private consultants also figure on the list of presenters. Only one academic was a presenter on the agendas of the three conferences, which took place. In addition to the presenters, other technical personnel were in attendance, in proportions roughly similar to those above, judging from the one attendance list we were able to examine.

The publications philosophies of the two major industrial players in cellular is another factor that impacted the publication of articles documenting the intellectual development of the cellular phone. Motorola discouraged its staff from publishing in the public literature, and at the time did not have a regular in-house technology organ. Bell Labs had a more open publication policy, and encouraged its staff to publish and be active in their technical communities. The Bell System Technical Journal was the vehicle for many publications by Bell Labs scientists and engineers.

We conclude from our review of the available literature (primarily Bell Labs and a few other articles) that the knowledge base pertinent to the cellular phone was developed and transmitted almost exclusively within Bell Labs until the late 1960s/early 1970s when the details of their cellular system design became public with AT&T's 1971 proposal to the FCC. While the Bell Labs scientists and engineers who were working on cellular clearly tapped the general knowledge base by following the relevant public literature, including the few radio communications articles by academics, they have told us that they found little there which had a direct bearing on their cellular work.

It also clear from our interviews that the scientists and engineers at Motorola, who in those early days were then keenly interested in the design and manufacture of FM mobile radio equipment, were aware of the possibilities for cellular telephone systems, and followed developments closely. Motorola designed and manufactured the majority of the mobile hardware for AT&T mobile radio systems, and also designed and manufactured equipment for some of the earlier cellular trials. It should be remembered that during the era in which cellular telephony was being developed, AT&T still had a near monopoly on telephone service, and it was implicitly assumed by most observers that if cellular systems were to be fielded, that AT&T would field them. When AT&T surfaced its cellular concept in the 1971 proposal to the FCC, it included considerable capacity for dispatch systems, heretofore nearly the sole preserve of private operators using Motorola equipment. This was perceived by Motorola as an attempt by the Bell System to dip into their lunch box, and was a major incentive for Motorola's own cellular development efforts.

Motorola's own development of a cellular systems concept became public in 1973, with their presentation of "The DYNA T.A.C Concept" to the FCC. The Motorola concept for a cellular system was again surfaced in a paper ("A System Plan for a 900-MHz Portable Radio Telephone") by James J. Mikulski of Motorola, published in IEEE Transactions on Vehicular Technology in 1977. The paper outlined Motorola's thinking on how a system focused on portable mobile units might be laid out, and advanced the concept of cell sectoring as a way to reduce interference. It also showed that Motorola engineers had concluded that geographic location was unnecessary, and that signal strength was the best criterion for control and handoff decisions.
 

V. The Role of Intellectual Property Rights

Formal intellectual property protection was not an important factor in the development of the analog cellular phone (although it would become more so as development moved toward digital cellular). Transfers of intellectual property through licensing or other techniques were not a significant element of the cellular phone development. Bell labs did not often patent systems (in contrast to hardware) actively in the 1950s and 1960s. However, they did begin to file for a number of patents on the cellular concept in the 1970s, beginning with Patent 3663762, Mobile Communication System, invented by Amos E. Joel, Jr., of which Bell Labs was the assignee. The patent, which dealt with the switching hardware needed for handoff, was filed on December 21, 1970 and issued on May 16, 1972.

Prior to the 1984 consent decree, AT&T enjoyed a near monopoly on telephone service, and it was the only firm capable of implementing new developments. Thus, patenting was a relatively unimportant part of their corporate strategy in areas related to telephony. Rather, their primary intellectual property interest (according to one of our interviewees) was to block others from patenting important technical developments, which could be important to any aspect of telephony. For this reason, they sometimes published new developments (usually in the BSTJ) as a strategic measure to deter others from attempting to patent similar technologies.
 

VI. Contributions of NSF and other Publicly Funded Research

As is evident from much of the preceding, there was little publicly funded research that directly supported the development of the cellular telephone. Our interviews indicate that a few of the cellular pioneers received public support for their graduate education, primarily from the G.I. Bill and defense programs. None of our interviewees could identify any person who might have received any sort of NSF support for their education, except for Don Cox, who had a one-year NSF traineeship (which was not subject oriented, he recalled) at Stanford before going on to Bell Labs, where he became one of the key researchers in the group working on the development of the cellular telephone.

Although Bell Lab researchers drew on the nation's knowledge base, they have told us that little of the public knowledge being generated by universities and federal research institutions had any application, except in a general way, to the development of cellular's intrinsic technologies. Motorola researchers told us the same thing. Nonetheless, the National Science Foundation was active in a very modest way in research related to urban radio communications. Searches of the NSF Awards Database reveal that during the 1960s, the Foundation made four awards related to radio communications (totaling $195,200), and during the 1970s, made 15 awards in related areas (totaling $1,061,406). The Bell Labs researchers with whom we spoke were aware of some of the more relevant pieces of this work, but told us that it was focused on areas not pertinent to their cellular development activities.

The dearth of academic research in mobile radio was confirmed during an interview (for another project) at Columbia University's Center for Telecommunications Research, an NSF Engineering Research Center (ERC). A senior Center faculty member told SRI, that "until 1984 or so, 99+% of telecommunications research was conducted at Bell Labs. There wasn't anything going on at universities. About 78% of our (The ERC's) graduates went to Bell Labs..."]
 

VII. Digital Cellular--A Technological Epilog

The cellular phone that had its commercial debut in the early 1980s was not the ultimate solution to the problem of rising demand for mobile telephone service - it was only a step along the way. As noted above, as early as the late 1980s, cellular service was becoming congested in a few large cities, particularly New York and Los Angeles. Cell splitting turned out to be only a partial solution to growing demand. The next significant step in the growth of mobile telephony came from a series of technological developments under the rubric of digital cellular.

Digital cellular systems are divided into two general (and technologically incompatible) families, time division and code division multiple access. The time division multiple access systems include Time Division Multiple Access or TDMA, which was developed by Bell Labs and is being introduced in a number of cellular systems (mostly wireline systems) in the U.S.; and the Global System for Mobile Communications or GSM, a time division system developed as a pan-European standard and now in use in Europe and many non-European countries. The other major digital family is code division multiple access or CDMA, of which the best known is the proprietary system developed by Qualcomm, Inc., and for which Qualcomm's CEO, Irwin M. Jacobs, was awarded the U.S. Medal of Technology in 1994.

CDMA is a cellular operating system based on spread spectrum techniques. Spread spectrum is a technique (first developed for use by the military) which involves 'spreading' many individual transmissions (i.e., conversations) across the entire bandwidth of the system. There are a number of techniques to accomplish this, of which the two most important are fast frequency hopping and direct sequencing. In fast frequency hopping, the frequency of a given transmission is changed many times per second (to other frequencies within the system's bandwidth) according to a set of directions or code known to both the transmitter and receiver. Direct sequencing (the basis of Qualcomm's CDMA) involves using a formula to expand a given transmission in a certain way so that it occupies the entire bandwidth, and again, the code to the formula is known to both the transmitter and receiver.

In addition to the different transmission technologies, all digital systems require some way to digitize (and usually compress) conversations, and there are a number of such systems in use, with new techniques being developed all the time in universities, industry and elsewhere. The most recent development in mobile communications is the recently commercialized Personal Communications Systems (PCS). Whether PCS represents anything technologically new is open to debate. One observer has described PCS as "cellular in a different frequency band. The introduction of PCS services was made possible by new frequency allocations for land mobile radio. Most PCS services use digital techniques, primarily GSM and CDMA. Some are "cellular" in that they are based on classic cellular architecture, and some are more like older mobile dispatch systems with a few powerful transmitters and multiple receiver sites within a given market. PCS systems frequently offer additional features and services, such as paging, which are not available from cellular providers. However, it is only a matter of time until cellular providers begin to match the offerings of competing PCS systems.

The development of digital cellular, especially GSM and CDMA drew on a number of developments in academia and in firms founded and operated by former academics. As noted above, TDMA was developed, as was analog cellular, primarily in Bell Labs. We conducted a brief review of NSF awards and conducted telephone and email interviews, and could find no direct NSF involvement in the development of either of the current U.S. commercial digital systems, TDMA or CDMA. However, there were a number of academics with NSF funding who did work that was related to CDMA, and possibly to GSM, the European developed system.

George Turin of the University of California at Berkeley, told us that he believes that some of the work on "Problems of Urban Radio Communication" he did with NSF support in the mid-1970s was used by the European engineers who developed the GSM standard. Furthermore, NSF support enabled George Cooper to pursue his work on spread spectrum techniques at Purdue in the 1970s, work that eventually led to several patents in spread spectrum techniques and other areas related to mobile radio. Cooper's spread spectrum work was focused on the development of fast frequency hop techniques which, to date, have not been employed in commercial cellular systems. However, the work certainly contributed to the knowledge base in urban mobile communications, particularly spread spectrum techniques, which were seen as a promising technological answer to adding additional capacity to mobile radio systems. David Goodman of Bell Labs also published his work on a fast frequency hopping system plan.

The Foundation had another indirect influence on the infrastructure supporting today's wireless industry. In 1985, Theodore Rappaport was a graduate student at one of NSF's first ERCs, Purdue's Center for Intelligent Manufacturing, where his doctorate work focused on high frequency propagation and systems design for use in factory buildings. From this early ERC experience to his Presidential Faculty Fellowship, the Foundation has been key to the development of Rappaport's academic career and the development of his mobile radio program at Virginia Tech. His idea for the structure of the Virginia Tech's Mobile and Portable Research Group (MPRG) with its industrial associates was modeled after NSF's ERC concept. He told us that the Presidential Faculty Fellowship he won in 1992 led to early tenure and allowed him to write his mobile radio textbook, Wireless Communications: Principles and Practice (1996), the first of its kind. Since its founding, the MPRG has graduated hundreds of electrical engineers who have gone on to work in the wireless industry. Virginia Tech also is home for the Center for Wireless Technology (CWT) of which Rappaport was a co-founder in 1993, specializing in networking, satellite communications, and antennas, while MPRG focuses on propagation and antennas, system design, simulation, Digital Signal Processing, and CDMA technology.

Prior to the establishment of MPRG and other new centers such as WINLAB at Rutgers, founded by David Goodman, formerly of Bell Labs (an NSF funded I/UCRC), there were virtually no academic programs to prepare students to work with wireless and/or cellular technology. Prior to their establishment, the wireless companies (primarily Bell and Motorola) hired electrical engineers and gave them on-the-job-training in wireless. However, as the wireless industry grew, the on-the-job-training option was not feasible for the large number of smaller firms entering the wireless business. Thus these new academic programs have been very successful and have provided a heretofore nonexistent source of trained wireless engineers.
 

VIII. NSF Managerial Strategies

In the first two decades of the NSF's existence (by which time the engineering challenges of the analog cellular phone had largely been surmounted), response to unsolicited proposals was the primary management strategy. This strategy had limited ability to select or encourage any area of investigation over another, and, as one of our interviewees noted, tended to favor established lines of investigation familiar to peer reviewers as opposed to radical departures. Particularly since there were only a very few academic researchers working in fields related to urban mobile radio, one would expect very few unsolicited proposals and awards in this area. As noted above, we could find only four awards in areas related to mobile radio in the 1960s, and 15 in the 1970s. As NSF's management strategy evolved, the sponsorship of focused meetings and workshops, first among academics, and later among academics along with industrial researchers began to become more common. We could not identify any awards for a meeting related to mobile radio.
 

IX. Conclusion

NSF Role

Despite this case's evidence indicating that NSF played little or no direct role in the development of the cellular phone, it would be erroneous to conclude that NSF "missed the boat" on cellular. In fact, NSF did the right thing in the early days of cellular development staying out of mobile radio research, and made the right move to change its involvement years later when the market situation had changed. It is easy to overlook the fact that the technological and commercial environment for telecommunications in the 1950s, 60s, and 70s was very different than it is today. In those earlier decades, AT&T was the near-monopoly provider of virtually all telecommunications services. In addition, it owned and operated one of the nation's premier research facilities, Bell Laboratories. Besides conducting development activities for telecommunications systems to be fielded by AT&T, Bell Labs performed research which contributed to the nation's store of scientific knowledge in telecommunications as well as other important areas of science (the Nobel-winning discovery of the cosmic microwave background radiation comes immediately to mind). Not only was AT&T, through Bell Labs, the organization best positioned to conduct telecommunications research, it was virtually the only organization that could implement the results.

Had NSF made significant research investments in areas related to telecommunications in the 1950s, 60s, and 70s, observers of the day might well have asked why the Foundation was trying to compete with Bell Labs in an area already well addressed by the latter institution. Furthermore, Bell Labs, as well as the other firms involved in telecommunications, had usually drawn their science and engineering staff from among graduates of traditional university science and engineering programs and provided them with the additional education, training, and the on-the-job experience required to work in specialized research areas. Thus, at that time, there was no strong demand for new graduates with specialized degrees in wireless or mobile radio. In the absence of such demand, it is not surprising that there were very few academics working in those areas. Neither is it surprising that few academics proposed work in telecommunications to NSF, nor that few proposals were funded.

However, in the late 1970s and 1980s, a number of things changed. The importance of many new types of telecommunications technologies and services and their critical national role as supporting technologies became obvious. In addition, the Bell System was broken up, and Bell's monopoly on most forms of telecommunications was lifted. Finally, many small telecommunications providers began to grow and enter new areas, such as cellular, and even more new telecommunications companies were being formed. Many of these companies became very successful and grew quite large. But there was very little public research or human capital to support these new firms that generally could not, as Bell Labs, Motorola, and other large firms had, train their own specialized manpower in-house. Academia, supported by NSF and others, began to respond.

As noted above, in the 1970s, as the cellular idea began to look promising, industry researchers began to meet more frequently, first at the specialized seminars held in Boulder, Colorado, and later at expanded sessions during mainstream telecommunications conferences. There they came into contact with the growing number of academic researchers in the field. In the 1980s and 1990s, a few cellular and wireless researchers from Bell Labs and Motorola left these companies for academia to develop programs in wireless communications or to found start-up companies to fill new technology niches in cellular and mobile communications. Academia began to pay more attention to the needs of this growing community, with new programs and centers focused on wireless, mobile radio and related areas. These programs were founded by retired veterans from the telecommunications industry as well as by the growing number of young academics who did their graduate work in telecommunications areas.

As the market for increased academic work in wireless, urban mobile radio, and other new areas of telecommunications grew, NSF began to change its awards profile. The number of awards related to telecommunications increased, and the Foundation began to pay more attention to these fields as several university programs in wireless entered the ranks of NSF supported centers. As the following chart indicates, NSF support for the broad field of telecommunications grew substantially during the 1990's, reaching a peak in 1994.
 

Other Public Support

Other federal agencies played a more prominent role in supporting telecommunications R&D and thus had a tangential impact on the development of cellular. The military services and defense agencies in particular invested substantial sums in telecommunications research. A good deal of this money went to academic institutions, enabling them to support the education of graduate students. Several of the cellular pioneers we interviewed noted that they had undertaken some of their graduate work with support from defense sector funds. For some of the earlier cellular researchers, the G.I. Bill was also an important source of support, as it was for many other scientists and engineers of that era.
 

From Fundamental Research to Flip Phones

Although some basic research, particularly research on the propagation characteristics of higher frequency bands, was essential to the development of cellular, most of the development effort consisted of continuous improvement in existing communications products and technologies and then the fusion of a number of existing technologies to produce a substantially improved product. At the time the cellular concept first surfaced in print at Bell Labs in 1947, the major technological challenges and a series of potential solutions were identified. Although many of these solutions could not be achieved at the time and in some cases would await developments in other fields (integrated circuits, for example), no major breakthroughs in knowledge were required to kick the development of cellular into high gear. In fact, that kick ultimately came not from the laboratory, but from events in the regulatory and business sectors. Not until the late 1960s, when the FCC sent strong signals that they were ready to make a significant allocation of frequencies for mobile radio, did the cellular development program at Bell Labs really get under way.
 

X. References
 

• Calhoun, George, Digital Cellular Radio, Artech House, Norwood, MA, 1988.
• Grant, August E., ed., Communication Technology Update, 3rd Edition, Butterworth-Heinemann, Newton, MA, 1994.
• Jakes, William C., Microwave Mobile Communications. New York: Wiley, 1974.
• MacDonald, V. H., "The Cellular Concept," Bell System Technical Journal, vol. 58, no. 1, January, 1979.
• Motorola Cellular Telecommunications Milestones, Document from Motorola Corporation internal Web server.
• Motorola Cellular Telecommunications Milestones, Document from Motorola Corporation internal Web server.
• O'Neill, E. F., ed., "A History of Engineering and Science in the Bell System: Transmission Technology (1925-1975)," AT&T Bell Laboratories, 1985.
• Okumura, Y. "Field Strength and its Variability in VHF and UHF Land Mobile Service," Rev. Elec. Comm. Lab. September 1968.
• Rappaport, Theodore S., Wireless Communications, Principles and Practice, Prentice Hall, Upper Saddle River, NJ, 1996.
• Ring, D.H., "Mobile Telephony - Wide Area Coverage," Bell Laboratories Technical Memorandum, December 11, 1947

Introduction

Two experiments were undertaken in order to test the value of specialized analyses of bibliometric and patent data for the retrospective study of cases of engineering innovation. The bibliometric experiment involved the use of three annual Research Front Databases produced by the Institute for Scientific Information using its citation indexes for the years 1988-1990. The patent study, conducted by Mogee Research and Analysis Associates (MRAA) under subcontract to SRI, involved a number of analyses based on citation patterns among relevant patents. Both experiments proved successful in demonstrating the utility of the analyses for developing innovation case studies. They also provided information that was consistent with several findings of the study, including the fact that foreign technology played a relatively small role in the development of U.S. cellular telephony.

The ISI Research Front Databases are produced by a technique called co-citation clustering. A computer algorithm is used to identify highly cited papers in ISI's annual citation index database, including the Science Citation Index (SCI), Social Sciences Citation Index, and CompuMath Citation Index, the latter adding a number of computer science and mathematics journals to the set covered by the SCI. Highly cited papers are then clustered together on the basis of the frequency with which they appear together (i.e., are "co-cited") in the reference lists of papers indexed by ISI during the year of the database (current papers). The number of times such pairing takes place establishes the strength of the "co-citation" link used by the clustering algorithm. The typical result is about 8,000 clusters of highly cited papers that usually represent the underlying concepts, discoveries, or methods driving a research area that approximate what scientists view as their research specialty. That is to say, the cited papers usually represent important contributions to an ongoing problem area, and authors of the current papers represent the scientists in the peer group active in the "research front" of the specialty: hence the name Research Front Database. Papers indexed during the database or previous years by ISI (source papers), whether current citing papers or older cited papers defining the cluster, carry information about all of the authors, bibliographic citation, and the institution(s) with which the author(s) are affiliated.

Although the research leading to cell phone technology originated more than twenty years before the three databases available to SRI, the objective of this effort was to use the names of a number of the researchers involved to identify:

• older cited papers by these researchers, along with those linked to them by co-citation, that continued to be recognized as part of the basis of ongoing work in 1988-90; and
• authors of the associated current papers to determine the content of their current work and whether these authors were part of the network of cell phone developers identified by other means during the case study.

• identify a relatively large set of patents that, through cocitation linkages, can be shown to be related to the key cell phone patents;
• identify patent clusters that might provide some insight into inherent technological clusters within cellular telephony;
• within this larger set of related patents, determine the extent and nature of citations to the research literature as an indication of science-technology linkage;
• identify the contributions made to the innovation by foreign intellectual property.

Bibliometric Results

The bibliometric study was intended to test the viability of using co-citation data as a tool to augment knowledge collection during the construction of the engineering innovation cases. The case dealing with cellular telephony presented a challenge to the experiment in that the original technology was almost entirely developed by industrial research efforts and therefore had very limited visibility in terms of the open published literature. As the discussion of "The Flows of Knowledge" (above) notes, few academics were involved in research in the field, while much of the Bell Labs work appeared in internal Technical Memoranda. The other major player, Motorola, has a policy of very tightly controlled and limited publication. The search of the Research Front Databases was therefore based on key researchers involved in the follow-on technology of "digital cellular communications" as they had been identified during the evolution of the case study.

Of the names selected for input in searches of the three 1988, 1989, and 1990 Research Front Databases, only five were identified as either a cited or citing author. Table A-1 shows the search names found, their affiliations and the total number of cited papers, if any, for each. A total of 55 papers were identified with these authors, by far the most having been authored by Kailath, a mathematician involved in the study of signal processing. Kailath had contributed 17 cited papers, and Sundberg one: the other papers were citing papers indexed during one of the three database years.
 
 

Upon examination of the 46 clusters associated with one or more of the identified papers, one proved to be of particular interest. A 1990 cluster dealing with applications and signal capture in mobile radio channels was defined by four cited publications, including the 1974 compendium by William C. Jakes, Microwave Mobile Commu-nications referred to in the discussion of "The Flow of Knowledge. The others were a 1982 book by Lee, Mobile Communication, a 1987 paper in Proceedings of the IEEE by Cox, and a 1979 paper on "Cellular Concept" by V.H. MacDonald published in the Bell System Technical Journal. Cox, Lee, and another individual interviewed for the case study, P.T. Porter of Bellcore, had also published citing papers in the 1990 cluster. The cluster had thereby identified a number of the cellular pioneers and others selected by other means for interviews, suggesting that searches of this type would prove a valuable way of validating and extending the range of knowledge developed at the inception of case studies of this type.

This conclusion was strengthened by a number of other clusters. One cluster was a 1989 research area dealing with signal modulation in digital communications. Others were identified through citation links among clusters. Because the citing papers may refer to the defining papers of more than one cluster, the Research Front Database is able to group such clusters into larger sets of related research areas. One such set of five 1990 clusters included the Jakes et al. cluster discussed above and, collectively, included papers by all five of the authors in Table A-1. Based on the somewhat limited treatment of digital cellular telephony included in the case study, it was concluded that these bibliometric data would represent an excellent starting point for a case study on that topic, which has involved far more international and university-based research than did the original cell phone technology.

The bibliometric experiment therefore led to the conclusion that the use of such data to "map" a case study topic area would represent a valuable tool to augment the process of developing an understanding of the research related to a given innovation. In addition, it represents a useful way of validating and extending the understanding of the players and potential interviews needed to develop the case study. A limitation is that the technique is much more appropriate to science-based innovations or to examining the science that underlies a selected technology.
 

Patent Analysis Results

The analysis began with ten key patents. SRI identified these key patents by first carrying out a search of U.S. patents available in electronic databases (1971-present). The search for cellular telephony and its related concepts yielded 716 patents. These patents directly referred to cellular telephony or its technological synonyms. Within the set of 716 patents, the patents were ranked according to how frequently patents within the set cited each other. A list was compiled containing the ten most highly cited cellular telephony patents (Table A-2).
 
 

Following this, MRAA retrieved the patents referenced by the 10 key patents, then they retrieved the patents referenced by those patents, and so on, until no additional cited patents could be found in the time frame covered by the databases used. The resulting set of 1,632 patents is the patent citation network; it consists of the 10 key patents and the 1,622 cited patents MRAA were able to retrieve. MRAA used three different databases to trace the history of cell phone patenting through patent citations. MRAA's own Patent Citation Analysis Database (PCAD) was used for patents issued since January 1, 1975. The on-line database USPM, published by Derwent Publications, and IBM's Patent Server Database, available on the Internet, were used for patents issued between 1971 and 1974.

The patent citation tracing began with the ten key patents. The 8 of these patents that were issued after 1975 were searched in PCAD to find all of the patents they cite. These patents cited 55 patents that were divided into three groups. Those of the 55 patents that were issued after 1975 were again searched in PCAD to identify the patents they cited. Those issued between 1971and 1974 were put into a file to be searched in USPM or IBM's database (along with the two patents from the original list of key patents that were issued before 1975). Patents issued before 1971 were put into a third file. Citation data are not available in electronic form for patents issued before 1971, so these patents could not be traced back any further.

MRAA went through six iterations of patent citation searching in PCAD, at which point no new cited patents were found that had issued after 1975. Beginning with the 8 key patents, MRAA found that they had cited 55 patents. Those issued since 1975 in turn cited115 patents, of which those issued since 1975 cited 165 patents, which cited 117 patents, which cited 79 patents, which cited 11 patents. With each iteration MRAA added the cited patents with issue dates between 1971 and 1974 to the list to be searched in USPM or IBM's database, and added those with issue dates before 1971 to the list that could not be traced back further. There was considerable overlap among the patent citations found in PCAD, so that when MRAA finished with PCAD the result was a network of 413 patents, of which 320 were issued before 1975. For each of the patents in the network issued after 1975, data were available on their patent numbers, assignee, inventors, the patents they cite and the patents that cited them.

MRAA went through the same process in USPM and IBM's database for the patents issued between 1971 and 1974. Each patent was searched to find the patents it referenced. Any of the cited patents that had issue dates between 1971 and 1974 were again searched to get the patents they referenced, while those issued before 1971 were put into a file of patents that could not be traced back further. Three iterations using these databases were carried out before no new patents issued after 1971 were found. For most of the patents issued between 1971 and 1974 MRAA obtained patent numbers, assignees, inventors, and citation data. Some data from 1971 were found to be incomplete, however, so for some patents issued in 1971 data are missing. While data on patents issued before 1971 are not available in electronic form, USPM and IBM's database give the last name of the first inventor on patents listed in the citation field of the patents they cover. Thus for patents issued before 1971, MRAA in most cases identified the last name of an inventor.

MRAA ended up with 1,632 U.S. patents in the final network. Of these, 93 have issue dates since 1975, 402 have issue dates between 1971 and 1974, and 1,137 were issued before 1971. They calculated the number of times each of these patents have been cited by other patents within the network (in-set citations). They also searched all 1,632 patents back through PCAD to get the number of times that each of these patents has been cited by any U.S. patents between 1975 and the present.

MRAA ranked patents by the number of in-set citations and citations since 1975 that they have received. They also attempted to rank companies and inventors by citations. Because MRAA only had assignee data on patents issued since 1971, the citation data accumulated for over two-thirds of the patents in MRAA's network could not be matched to an assignee. For the data they were able match to assignees, a relatively small number of large companies accumulated most of the citations. It is unknown whether these companies would continue to rank at the top if the remaining data could be matched. The ranking of inventors was somewhat more difficult. MRAA had some inventor data for almost all the patents in their network, but the data are often incomplete. That is, for some patents MRAA had full inventor names and addresses, while for others they had a last name for one inventor. MRAA did some clean up of inventor names; however, they decided to err on the side of caution. If a last name was unique, they unified it with a record that had a complete name with that same last name, but a record that just had the name "Anderson" was not unified with another record that had "Dean T. Anderson". This inability to unify common last names, and the fact that most of the patents in MRAA's network have a last name for just one inventor on the patent, means that the citations for an unknown number of inventors are undercounted.

The final network of 1,632 patents were used in three analyses:

• a co-citation cluster analysis
• a literature reference analysis
• a foreign reference analysis.

The cluster analysis provided an opportunity to develop a graphic "map" of cellular technology as represented by the patent network. Like the bibliometric experiment, this involved the technique of co-citation analysis, here defined as the number of times two patents were cited together by a subsequent patent in the network.The analysis was designed to identify patent clusters that can provide some insight into inherent technological clusters within cellular telephony. The analysis began with the 1,632 patents in the network and identified all patents issued between 1971 and the present that cited them. Using this larger set of 10,416 patents, a smaller set was identified that had been cited at least three times-an indication of their significance and the equivalent of the bibliometric clustering's first step of identifying highly cited papers. This yielded 166 patents in 28 clusters with the clusters ranging in size from 3 to 29 patents. Most of the patents in the citation network are quite old and were not receiving many citations after 1975. For this reason the co-citation thresholds used to get clusters to form were quite low and the resulting clusters may be somewhat noisy. A few clusters point to some distinct technological subclusters such as error control, phase lock loop, and data encoding. Others point to interesting linkages such as the pointer to paging. Eight of the ten key patents identified at the outset of the analysis are in the same cluster, which is shown in Figure A-1, with detailed information about each patent shown in Table A-3.
 

Notes to Table A-3:

Despite such distortions, the map confirms the finding that AT&T and Motorola were the only significant players in the development of the technology, Motorola having seven and Bell four of the patents showing on the map (the only unassigned patent was invented by Cox, an AT&T employee). Three patents are assigned individually to other companies: Martin, Harris, and NTT. Bell's fundamental underlying patent lies at the center of the map, which can essentially be divided into four "regions." The Basic Cell System technology lies at the center with the exception of the patent that made Harris a brief player in the technology. Cell Sectoring and Splitting lies just below it with one Bell and one Motorola patent. To the left are patents dealing with Handoff and System Control (and two peripheral Motorola patents, one of which represents emerging digital technology). On the right side are three patents concerned with the technology of wireless data transmission.

The mapping technique is highly experimental and relatively crude. The co-citation links used for the matrix input for MDscaling were simple raw, unnormalized links. Nor was any effort undertaken to account for the fact that the network contains patents from many different years, so the strength of links between older patents is affected by the longer time period in which they accumulated. However, despite the distortions caused by forcing the representation into a two-dimensional presentation, the map provides a useful overview of the patented technologies underlying cellular telephony, including an intimation of the digital technology to come. We conclude that the technique merits further development and that it can contribute to the understanding of the technologies relevant to innovation case studies.
 

Literature Reference Analysis

This analysis was one measure of the extent to which there was significant direct academic intellectual content involved in cellular telephony. From the patent network of 1,632 patents, the analysis used only the 472 patents that were issued since 1971 and could be analyzed electronically. Of the 472 patents in this set, 45 (less than 10%) had references to scientific literature. Only two pieces of literature were cited more than once: the Bell Labs "High Capacity Mobile Telephone Technical Report" and C. Kramer, "A Low-Frequency Pseudo-Random Noise Generator." These results provide one indication that the direct academic intellectual content of cellular telephony is low. However, the ability to identify linkages to fundamental research through patent citations is hotly debated for a number of reasons. Among the most important are that practices of citing research literature may vary across technical fields and, within fields, by individual patent examiner. Nonetheless, we can compare the citation rate for these cell phone patents with similar rates in other technical fields and observe that the linkages to fundamental science in these patents, as measured by citations to research literature, is relatively low.
 

Foreign Reference Analysis

This analysis was designed to determine if there was significant foreign intellectual property involved in cellular telephony. The analysis also used the 472 patents that cited the ten key patents. Among this set, only 23 patents contain references to foreign patents (less than 5% of the total in the set). Of the foreign patents cited, only one foreign patent was cited more than once. These data support the conclusion that the foreign intellectual property content of cellular telephony is low.
 

Conclusion

Both the bibliometric and patent analyses proved to be useful adjuncts to the case study. The use of bibliometric co-citation clusters provides insight concerning the research that underlies some technologies, especially where there is a substantial academic element-of particular interest to NSF in this series of case studies. It also aids in identifying and validating the network of players who represent prospectively valuable sources of information in developing the case study.

The most interesting and promising of the patent analyses in this case was, again, the co-citation clustering experiment. While the bibliometric technique focuses on research, the patent clustering provided a useful overview of the technologies for a case in which these had largely been developed by industry and lacked an extensive background of university-based research. The analysis of patent citations to the scientific literature confirmed the low content of such research, while the analysis of foreign patent citations reflected the highly concentrated nature of the development of the technology by two U.S. companies, AT&T and Motorola. The two techniques are thus complementary and, together, can contribute usefully to the study of how innovations evolve.