INTRODUCTION BY THE EDITOR - WINLABnarayan/Course/Wireless... · 2019-07-25 · cell phone manufacturers create the new radio designs these phones would require. This concession gained

IEEE Communications Magazine • September 201014 0163-6804/10/$25.00 © 2010 IEEE

INTRODUCTION

I have attempted in this brief history tocapture the rich mix of technical andpolitical challenges that shaped thedevelopment of the Advanced MobilePhone System (AMPS), the first cellu-lar system in the United States, between1971 and 1983. Earlier work wasdescribed in an article written for thiscolumn by Joel Engel [1]. I have alsowritten a more detailed and personalview of these events within a manuscripttitled Cellular Dreams and CordlessNightmares; Life at Bell Laboratories inInteresting Times, which is available fordownload from http://www.winlab.rut-gers.edu/~frenkiel/dreams [2].

The cellular concept emerged atBell Labs in the late 1940s, in propos-als by Rae Young and Douglas Ring.Some modest additional studies wereconducted over the next two decades,including a conceptual system plan thatPhil Porter and I produced in 1966.The attraction of cellular systems wastheir ability to reuse channels manytimes in a local area, and to provideservice over very large areas with low-power radios, but a cost-effective cellu-lar system required a large number ofchannels. Channels for all radio sys-tems were in very short supply until themid-1950s, when advances in radiodesign made frequencies up to about 1GHz practical for consumer applica-tions. This essentially doubled the avail-able spectrum and led to competingproposals for a variety of purposes,including AT&T’s cellular system and

an expansion of TV broadcasting. Aftersome deliberation the FCC used thishuge block of new spectrum to createmore than 80 new TV channels. TVwas a relatively new and very popularmedium at the time, and there was anexpectation that this expansion wouldlaunch an era of cultural and educa-tional programming.

The results of that experiment fellshort of expectations, and in 1968 theFCC opened an “inquiry and proposedrule-making” to consider the reassign-ment of 14 of those UHF TV channels,near 900 MHz, for mobile telephoneand private mobile radio systems. Thiswould produce well over 1000 newmobile radio channels, and becausethese channels were considered to be alimited and valuable national resource,the FCC added the expectation thattheir inquiry would lead to more “spec-trum efficient” systems.

The cellular system met the criterionof spectrum efficiency very well, butearly plans for these systems had beenlimited in scope. Little was known aboutradio performance at 900 MHz, forexample, or about the cost and ultimatecapacity of a practical cellular system.To fill in these details, a large team wasassembled at Bell Labs to prepare a“feasibility study and system plan” forthe FCC. The core of this activity wasJoel Engel’s cellular systems engineeringgroup under Rae Young, to which PhilPorter and I moved in 1968, and a cellu-lar radio development department underBob Mattingly. The latter were radarexperts who had recently returned from

a long assignment on Kwajelein Island inthe Pacific, and included many whowould be key contributors to the AMPSproject. Radio research engineers underBill Jakes contributed new ideas in radiopropagation theory and receiver diversitythat Jakes would later compile in a land-mark text. [3] Zack Fleur, representingswitching systems engineering, providedexpertise on the switching systems thatwould be modified to become AMPSsystem controllers.

The resulting cellular plan was deliv-ered to the FCC on December 20, 1971[4], and AT&T committed its vastresources to the creation of a nation-wide cellular network. In anticipation ofFCC approval (and perhaps to encour-age that approval) AT&T immediatelylaunched a full-scale development pro-ject, estimating that cellular servicecould be offered within about five years.Sadly, this prediction proved overlyoptimistic. The development wentsmoothly, but intense controversiesover the role to be played by the AT&Tmonopoly would delay commercial ser-vice until 1983.

At that time there were many smallindependent telephone companies, butthe AT&T monopoly (often called theBell System) provided more than 80percent of the local and long-distancetelephone service in the United States.Moreover, most of the equipment forthe nationwide telephone network wasdesigned by Bell Labs and built byWestern Electric (AT&T’s manufactur-ing arm). In contrast, the mobile tele-

HISTORY OF COMMUNICATIONSEDITED BY MISCHA SCHWARTZ

CREATING CELLULAR: A HISTORY OF THE AMPS PROJECT (1971–1983)RICHARD H. FRENKIEL

The article following, written by one of the lead engineers in theproject, describes the development of AMPS, the first cellular tele-phone system in the United States. As noted, the project ran forover 12 years and required the services of a “vast number” of engi-neers at AT&T’s Bell Labs. The hundreds of millions of peoplecurrently using cell phones, smart phones, and all the other mobilewireless equipment available in the United States, as well as every-where else in the world, often take these devices and the infra-structure behind them for granted. We tend to forget that, not toomany years ago, the spectral environment over which our mobilesystems operate on a regular basis was not well understood. Untilthe cellular concept was recognized, it wasn’t clear that manymobile phones could simultaneously share the same channels in alocal area. With the cellular concept agreed on, problems such astracking of cell phones and handoff from cell to cell had to beresolved. Computerized base stations had to be designed and test-ed. A myriad other important development and design issues and

problems had to be resolved. Dick Frenkiel very clearly describesthe various engineering problems that had to be solved, step bystep, sometimes quite painstakingly. He gives credit to the manyteams of Bell engineers who worked on this project. He also pointsout policy issues involving the FCC that had to be resolved. Com-petition for scarce radio spectral space was fierce. Competing engi-neering companies and telephone operators had to be assuaged.Getting from the initial concept to successful implementation of atrial system in Chicago took massive funding and assignment ofmultiple large teams of engineers by AT&T. But the job was finallyaccomplished. My own personal feeling is that this project, like anumber of other massive projects requiring hundreds of engineers,could only have been undertaken at that time by a company withthe resources of AT&T. But that’s another question for anther day.In the meantime, I urge you to read on through this excitingdescription of the genesis of cellular telephony in the United States.

—Mischa Schwartz

INTRODUCTION BY THE EDITOR

(Continued on page 16)

LYT-HISTORY-Frenkiel-September 8/24/10 1:44 PM Page 14

IEEE Communications Magazine • September 201016

phone and private mobile radio mar-kets were relatively open and competi-tive. Small businesses called radiocommon carriers (RCCs) operated halfof the mobile telephone systems in theUnited States, and a number of radiomanufacturers provided equipment forboth mobile telephone systems and pri-vate mobile systems.

Both the RCCs and the radio manu-facturers had reason to be concernedabout the cellular proposal. The RCCswere generally small businesses, andwould have difficulty installing andoperating complex and expensive cellu-lar systems, even if they were to getaccess to an adequate number of chan-nels. The radio manufacturers expectedto lose the market for mobile telephoneequipment to Western Electric, andalso worried that cellular service wouldweaken the demand for private systems.Their concerns were shared by the anti-trust division of the Department of Jus-tice, which had sought to limit AT&T’smonopoly power for most of the centu-ry. The stage was set for a bitter andlong-term confrontation.

In an attempt to lessen the fears ofits monopoly power and gain wider sup-port for its proposal, AT&T announcedthat it would not manufacture cellphones, and that Bell Labs would helpcell phone manufacturers create the newradio designs these phones wouldrequire. This concession gained the sup-port of some manufacturers, but impor-tant concerns remained. There might bea large new market for cell phones, but astandardized cell phone would attractnew international competition and giveAT&T great buying power. The high-margin market for private, custom-designed, turnkey systems might bereplaced by a low-margin market for cellphones. The largest of the radio manu-facturers, Motorola, would remainstrongly opposed to the AT&T proposal.

Among the many opponents ofAT&T’s cellular proposal, Motorolawas perhaps the most effective. Itsopposition was always energetic, oftencreative, and occasionally rather witty.An amusing audio simulation, forexample, contrasted a long and ram-bling telephone conversation with thebrief and businesslike transmissions ofa fleet dispatcher, to suggest that “effi-

ciency” had some non-technical dimen-sions. Other challenges involved com-plex calculations of spectrum efficiency,and led to prolonged and confusingpublic debates between Bell Labs andMotorola engineers that did little toresolve the issues and much to delaythe project.

THE PERVASIVE ISSUE OFSPECTRUM EFFICIENCY

Because base stations are very expen-sive, cellular system designers and oper-ators must always focus on spectrumefficiency — if only to maximize thenumber of subscribers that share thecost of each base station. The modernFCC practice of auctioning new spec-trum has resulted in some very highprices, which has only served to intensi-fy the economic incentive to use spec-trum efficiently. In the 1970s, however,radio spectrum was seen as a free butlimited national resource, to be allocat-ed by the FCC for the greatest publicgood. This cast a strong and almostmoralistic light on the issue of spectrumefficiency: it was the means by whichone could serve the greatest number ofpeople with any given amount of pre-cious spectrum.

Perhaps the most significant issue indetermining the spectrum efficiency ofa cellular system is the separationbetween cells that use the same chan-nels —the so-called reuse distance.This, in turn, determines the number ofcells requiring different channels, andthus the number of channels per cell.[5] Using an example based on AMPS,if one has 350 channels and can reusechannels at a distance of 4.6 cell radii(Fig. 1), the result is a 7-cell repeatingpattern with 50 channels per cell.

For an AMPS call blocking objectiveof 2 percent and an estimated trafficload of 90 s/subscriber/busy-hour, a cellwith 50 channels will serve about 1600subscribers. In a hexagonal grid of 1-micells, this represents a capacity of about600 subscribers/mi2. If twice this reuseseparation were required, however, a28-cell repeating pattern would result,and the same 1-mi cells would serveonly about 100 subscribers/mi2 (fewerchannels per cell with smaller, less effi-cient server groups). The capacity of acell (and a system) would be reduced by83 percent, and the cost per user wouldincrease sharply.

This calculation requires only a fewseconds and an Erlang B calculator, butwith limited propagation data and nooperating experience, our task in esti-

HISTORY OF COMMUNICATIONS

Figure 1. A 7-cell channel reuse pattern (from AT&T's 1971 proposal to theFCC).

D/R = 4.6N = 7 sets

D R

2

1

6

4

3

1

1

1

1

5

1

7

1

1


(Continued from page 14)


18



mating reuse distance was a complexand time-consuming process. We need-ed to set an objective for voice qualityand then determine what design param-eters would meet that objective. Thisrequired an experimental program tomap the relationship between subjectivevoice quality and objective signal-to-interference ratio, where both signaland interference could be fading at var-ious rates (i.e., for different vehiclespeeds). To conduct any such voicequality experiments we needed to speci-fy a particular radio design, and to con-vert these laboratory results to a reusedistance we needed propagation data at900 MHz from a variety of environ-ments.

A massive program of propagationmeasurements in typical U.S. environ-ments was carried out by a team led byGerry DiPiazza, one of those radarexperts from Mattingly’s departmentwho were so important to the AMPSproject. Gerry’s propagation modelswere then processed in computer simu-lations, to determine the signal-to-interference performance of differentcell grids. The largest of these simula-tions, called MultiCell, was designed byJim O’Brien and could simulate thou-sands of calls in a mature system con-figuration. For many years, Jim’sMultiCell simulation was our best win-dow into the reality of an operatingcellular system. It was used to study awide variety of issues, from the varia-tion of signal and interference during atypical call, to the performance oflocating algorithms, to the logistics ofcell splitting and the ultimate capacityof a system.

Inevitably, the design of those exper-iments and the interpretation of theirresults involved an element of subjectiv-ity. Methods and conclusions were oftenchallenged, leading to time-consumingpublic debates. Moreover, improve-ments in spectrum efficiency were adouble-edged sword. They demonstrat-ed the value of cellular systems but sup-ported the argument that such systemscould be built with fewer channels.

Despite a variety of such controver-sies, the system design gradually tookshape [6]. FM channels with 12-kHz fre-quency deviation were shown to providegood voice quality in an interference-limited environment. “Companding”(the compression of the dynamic rangeof a voice signal before transmissionand its compensating expansion afterreception) and the use of a “knee” inthe expander curve to reduce back-

ground noise were shown by GastonArredondo to improve voice quality sig-nificantly. Two-branch receiver diversi-ty, proposed and demonstrated by BillJakes’ researchers, was adopted at basestations to further improve radio per-formance.

Based on listening tests using theseadvances, it was determined that “90%good-to-excellent” voice quality wouldrequire a signal-to-interference ratio ofat least 17 dB over 90 percent of thecell. Based on propagation measure-ments and computer simulations, thisinitially implied a 12-cell reuse pattern,but Phil Porter’s proposal to use direc-tive antennas at base stations reducedinterference enough to allow a 7-cellreuse pattern.

LOCATING AND HANDOFFOf all the design issues created by cel-lular, none was more symbolic of thenew technology than the need to locatethe vehicle to a particular cell and to“hand off” a call to a new cell when thesubscriber’s location changed. The sig-nal and interference conditions couldchange very quickly for a fast-movingvehicle in a small cell, so the locatingprocess needed to be performed everyfew seconds for each call. The resultinghandoffs could be frequent and neededto be performed without annoying clicksor gaps in speech that would be objec-tionable to the subscriber.

The goal of the locating and handoffprocess was first seen as preventinginterference by constraining calls to usethe nearest base station. Using one ofthe first simulations of a mobile unit ina cell grid, however, Gary Ott demon-strated that the best approach was sim-ply to use whichever base station couldprovide the strongest signal. The result,of course, is that the actual service areasof cells are quite irregular, although wecontinue to draw them as idealhexagons. Jim O’Brien’s MultiCell sim-ulation was later used to select a mea-surement rate for sampling signal levelat nearby base stations, and a decisionrule for handoff that balanced signalquality and processing load.

Achieving a “clean” handoff withoutdropped calls was also a challengingdesign problem. Out-of-band signalingwas slow and unreliable, so it becamenecessary to interrupt the voice channelfor this function. The audio would bemuted during this operation, but it wasdetermined in voice tests that a gap ofmore than one-quarter second in theaudio path would be objectionable. Asignaling method called “blank andburst” was proposed by Reed Fisher, in

which the audio was “blanked” while abrief “burst” of heavily coded data wassent to the mobile. Reed was also theprimary designer of the fast frequencysynthesizer, which allowed the cellphone to change channels during thatsame brief period. Although AT&Twould not manufacture cell phones,Reed was a major contributor to cellphone technology. The central switchwould transfer the call to the new cellwhile the mobile was changing chan-nels, and the subscriber would hearonly a minor “click.”

SIGNALING AND CONTROLTo carry out functions such as call setupand handoff, the cellular system wouldneed to exchange a good deal of infor-mation between base stations andmobiles. A 10-kb/s digital signaling ratewas introduced for this purpose. Thiswas a high signaling rate for mobileradios in that period and would requireextensive testing for error performancein the cluttered, interference-limitedcellular environment.

The various signaling functions usedin the AMPS system presented a vari-ety of different coding problems. [7] Asdiscussed above, the blank-and-burstfunction for handoff was carried out onvoice channels. It required high relia-bility in a short time, to preventdropped calls while causing only a briefgap in conversation. Other channelswere only used for signaling purposes.Sets of one-way “paging” channels, forexample, were used to alert mobilesthroughout the system to incomingcalls. All the mobiles in the system thatwere not engaged in calls would be lis-tening for incoming calls, so a “falsepositive” rate of even 1 percent on thepaging channels would lead to hun-dreds of false responses to each incom-ing call and put a heavy load on thesystem processor.

There were also channels that wereused by the system to identify mobilesand exchange information such as tele-phone numbers for setting up calls.The “dial-then-send” operation foroutgoing calls that is now so familiarfor cell phones was another of PhilPorter’s innovations, to make the dial-ing operation less hurried (and there-fore safer) and to save the channeltime that would have been lost duringdialing. With a bit of additional inno-vation, this clever idea would enablemodern text messaging. One of theoriginal creators of the AMPS archi-tecture, Phil was responsible for someof its most novel features.






Compared to today’s complex digitalcellular systems, such innovations mayseem rather ordinary, but at the timewe were breaking new ground. A fewyears earlier, the dialing of a call from acar without operator assistance was abreakthrough. Now we were bringingmobile telephony into the modern ageof communications. This required apowerful and flexible system controllerthat could behave as a telephone cen-tral office, interface with base stations,and control the locating and handoffprocesses.

Fortunately for the AMPS project,the Electronic Switching System (ESS)became widely available in the late1960s. Pioneered by Bell Labs, this soft-ware-controlled switch offered a flexibleand cost-effective platform for addingnew cellular capabilities. It is difficult toimagine a successful AMPS projectwithout the ESS, or without the largeand talented team of Indian Hill soft-ware developers who converted the 1AESS into the AMPS Mobile TelephoneSwitching Office [8].

AMPS would also require a cellphone smart enough to play its part inthe choreography of call setup andhandoff, and once again a new technol-ogy presented itself. Early in the projectit was assumed that the cell phone logicwould be realized in complex and spe-cialized integrated circuits, but with theemergence of the microprocessor in theearly 1970s we inherited a powerful andflexible tool with which to implementcell phone logic. In retrospect, the ear-liest proposals for cellular in the 1940swere well ahead of their time. To thoseearly visionaries the radio technologymust have seemed almost within reach,but it would take another quarter cen-tury to achieve the computer technolo-gies that made a practical cellularsystem possible.

As a result of the size and complexityof the system, the AMPS base station(or “cell site”) [9] was different fromanything found in earlier systems. Itincorporated separate radios for locat-ing measurements and call setup, andmultiple frames of voice radios. Ampli-fiers of that period could not providethe power and linearity needed to com-bine many transmitters onto a singleantenna without creating spurious out-of-band signals, so transmitters wereamplified individually. Groups of 16transmitters were then combined withcircular arrays of tuned cavity filters thatwere affectionately called “radialengines” because of their resemblance

to World War II aircraft engines. Adata frame controlled overall base sta-tion operations, and a maintenance andtest frame monitored its health. Therewas a line supervision frame for incom-ing wireline connections to the switchand a power plant with battery backup.Even the monopole mast and antennaswere new designs. This massive develop-ment effort was the responsibility of alarge and talented team at our laborato-ry in Whippany, New Jersey, many ofwhom had arrived with Bob Mattinglyin 1968 and would remain through thelong project.

The design of the switch and basestations was separated between NewJersey and Illinois, and there were mul-tiple cell phone manufacturers as well.It was therefore important not only towrite clear functional and interfacespecifications (a role that involved manyof the systems engineers in my depart-ment), but also to make sure that thosespecifications were understood andaccepted across the project. This latterrole often fell to Stu Tartarone and PhilDiPiazza, who had a particularly broadgrasp of the system architecture andcould identify important new opportu-nities. Stu was the first to recognize, forexample, that improvements in mini-computer technology could be used tocreate a smarter base station within amore distributed and flexible controlstructure.

Selling such ideas across the entireAMPS team required diplomacy, perse-verance, and a willingness to travelquite a lot. Phil would continue this“commuter” role for many years, as the

system entered its trial phase and wasreadied for commercial service. Hewould later “commute” between NewJersey and Illinois, overseeing the finalsystem tests in Chicago, and wouldmake the decision that the first cellularsystem in the United States was readyfor commercial service.

Overall management of the AMPSproject was the responsibility of FrankBlecher, whose ability and energyproved equal to the enormity of thetask. During the field trial in Chicagohe would use every opportunity to per-form his own “system tests,” and webecame used to his early morning andlate night cellular calls from one of theChicago test vehicles. He kept themorale of our large and far-flung teamhigh through delays and disappoint-ments, and much of the credit for thesuccess of the project can be attributedto his leadership.

THE LOGISTICS OF CELL SPLITTINGThe key to the cost-effective startupand large final capacity of the cellularsystem is the process of cell splitting.In the AMPS system cells as large as5–10 mi in radius (depending on ter-rain) could be used at startup to mini-mize cost. Through the process of cellsplitting, these would gradually bereduced in size to increase systemcapacity. Because base stations arevery expensive, however, it is necessaryto retain existing base stations in thenew grid as cells split. Figure 2 showsthe splitting pattern used in AMPS.New cells are 1/4 of the area of theprevious cells, so each round of cellsplitting will increase the system capac-ity by a factor of four. The smallestcell size (and thus the ultimate systemcapacity) is limited by practical issuesof base station placement and frequen-cy of handoff. The smallest practicalcell was thought at the time to beabout 1 mi in radius, but much smaller“microcells” are used today.

Even with the original base stationsincorporated into the new cell grid,however, cell splitting presented somesignificant challenges. The centralregion of the system had to be coveredwith a continuous grid of the new small-er cells, and more than one-quarter ofthe channels would have to be movedinto the smaller cells simply to achievethe same capacity that had existedbefore the split. Since these channelswere no longer available in the nearbygrid of larger cells, the capacity of thelarge-cell grid would be reduced, andadditional cells would be forced to split.

Figure 2. Cell splitting that preserves theold base stations (+) in the new grid ofsmaller cells (o) (from AT&T's 1971proposal to the FCC).

+

+

+




Dozens of new base stations would haveto be installed and cut into service atthe same moment. Hundreds of radiosin the old base stations would have tobe moved and retuned. It was a costly,labor intensive process and promised tobe very difficult logistically.

While drawing such a configuration,it occurred to me that this massive dis-ruption was being created to achieve anincremental increase in system capacity.At that moment, the larger cells wereserving almost the entire trafficdemand. If we retained the existingcells as a continuous “underlaid” gridproviding almost all the needed capaci-ty, we could add a few “overlaid” small-er cells “here and there,” with only afew channels, to provide the smallamount of incremental capacity weneeded. Those few channels could con-tinue to be used at the original basestations for mobiles that were in theinner portion of the cell (in effect, for asmaller cell that was co-sited with thelarger cell). Calls that left these newisolated cells could be handed off to theunderlaid grid, which would still pro-vide geographically continuous cover-age. Over time, more new base stationswould be added, and more channelswould gradually be moved into thosenew base stations until the old grid wasfinally replaced, but the process wouldbe gradual and manageable. This elimi-nated the logistic problems of cell split-

ting and reduced the number of cellsthat were needed at most points ingrowth rather dramatically. Simulationsusing Jim O’Brien’s MultiCell simula-tion showed that the average number ofcells in a growing system (and thus theaverage system cost) were reduced bymore than 50 percent. That simple idealater became one of AT&T’s mostsought after patents in cross-licensingagreements [10].

SYSTEM TRIALS ANDCOMPATIBILITY STANDARDS

As the system design progressed towardcompletion, it became necessary todemonstrate that AMPS would provideboth the excellent service and the spec-trum efficiency that had been promised.This sort of demonstration would be afeature of any development program,but amid the political debates of thetime it led to new controversies anddelays. To demonstrate real service, wewould need a large coverage area, andto demonstrate the proper working ofthe system for the largest capacity wewould need to use small cells. Puttingthe two objectives together would cre-ate a trial with hundreds of cells, whichwas economically infeasible, so AT&Tproposed to separate the demonstrationinto two trials. The Chicago ServiceTrial would demonstrate real service ina startup configuration, with productionequipment and several thousand sub-scribers. A cellular testbed in Newark

would simulate operation in a few 1-micells, surrounded by six interfering cellsseveral miles away. The coverage mapsfor the Chicago and Newark trials areshown in Fig. 3.

Objections were raised to the pro-posed trials, in particular because theywould not demonstrate productionequipment in the smallest cells. TheFCC agreed with the objectors anddenied permission for the trials. AT&Tappealed, and the FCC eventuallyreversed their decision, grantingapproval for the Chicago and Newarktrials on March 10, 1977, but a full yearwas lost in the appeal process.

The FCC also granted approval toMotorola to operate a trial in the Wash-ington, DC/Baltimore area. A Motorolateam led by Marty Cooper had createdthe first truly portable handheld cellphone, called Dyna-TAC, and servicefor portable handsets would be thefocus of that trial. Dyna-TAC was largeand heavy by modern standards (aboutthe size and weight of a brick), but itrepresented a significant breakthroughin portability. It was a major step in theevolution of cellular from a telephonein a car to the truly “personal” commu-nication service we enjoy today.

Installation of the equipment andfacilities for AT&T’s Chicago trial[11–13] was the responsibility of a largeteam led by Jim Troe. The system used10 cells to cover 3000 mi2, with a switch-ing center at Oak Park and an opera-

22



Figure 3. Coverage maps for the Chicago and Newark trials.

Cellular Service Trial — Chicago Cellular testbed — Newark




tions center at Elmhurst where cellphones manufactured by Motorola, E.F. Johnson, and OKI were installed invehicles (the Chicago system was alsoused to prove-in the cell phone designsof several other manufacturers). OnJuly 1, 1978, the system began a 12-h/day equipment test with 25 test vehi-cles, and on December 20, 1978, itbegan full-time service to paying cus-tomers. It was a large-scale and fullyoperational cellular system, but in theabsence of a final FCC ruling it wouldremain a “trial” for five long years,capped at about 2000 customers, whilecommercial cellular systems were beingintroduced around the world.

The cellular testbed in Newark [14]was built by a team led by Gerry DiPi-azza. It used a heavily equipped testvan to test all the signaling and controlfunctions required for small-cell opera-tion while simultaneously gatheringdata on signal, interference, and audioquality in a densely populated urbancenter. The thousands of hours of datait provided were a final confirmationthat the promise of cellular would bekept.

Data from the two trials was com-piled by Ray Pennotti into a series of“90-day reports” to the FCC. [15] Overa period of several years, those reportsdemonstrated that Chicago customerswere happy with the service and thatsystem performance could be main-tained within the smallest cells. Wewere approaching the finish line, butthe trials continued, and the issue ofstandards remained.

The FCC had decided, by the late1970s, to split the proposed spectrumbetween cellular and private systems.The cellular spectrum would be furtherdivided between two competing sys-tems in each service area, one to beoperated by the local telephone compa-ny and the other by a private competi-tor. This final compromise, whichmirrored the pre-cellular licensing pro-cess, was proposed by Lou Weinberg ofAT&T’s Federal Regulatory organiza-tion. It made the field sufficiently com-petitive to satisfy the FCC, whileallowing each competitor sufficientchannels to design a cost-effective sys-tem. Lou would later go on to createthe AMPS corporation, a separate sub-sidiary created by AT&T to plan andimplement a nationwide cellular net-work, in the brief but euphoric periodbetween FCC approval and thebreakup of the AT&T monopoly.

By 1980, it was clear that the pro-

posed system would work as promised,but the FCC demanded agreement on asingle nationwide “compatibility stan-dard” that would allow any cell phoneto operate in any system. This work fellto an ad hoc committee of the Elec-tronics Industries Association. A draftwas prepared for the committee by TomWalker, based on the AMPS design (asdemonstrated in AT&T’s Chicago andNewark trials). After some modifica-tions within the committee, the pro-posed standard was submitted to theFCC. The record was complete, and onApril 9, 1981, the FCC issued a finalruling in their long inquiry. On October13, 1983, AT&T’s Chicago Service Trialbecame the first commercial cellularsystem in the United States, but theexpansion of cellular to other cities wasdelayed for several additional years by acumbersome and litigious licensing pro-cess.

POSTSCRIPTThe market, of course, turned out to bemuch larger than even our most optimisticpredictions, but AT&T would watch the firstdecade of that rapid growth from the side-lines. In 1984 the Justice Department’s longanti-trust case against AT&T finally endedwith the breakup of the AT&T monopoly.AT&T agreed to divest itself of its local tele-phone operating companies (which wouldoperate the cellular systems), and thus lostthe cellular networks it had fought to create.We had planted the seeds for a revolution-ary new service, but they would take rootand flourish in a transformed telecommuni-cations industry we had never imagined.

The Bell Labs engineers whodesigned AMPS would remain withAT&T, and would later become part ofthe spinoff to Lucent Technologies.Within Alcatel-Lucent, a few membersof our original team are still at work —designing complex digital cellular sys-tems that make AMPS seem rather sim-ple by comparison.

In this brief history I have been ableto name only a few of the hundreds ofBell Labs engineers who poured theircreativity and energy into that long pro-ject. A few others are named in the ref-erences and within the referencedpapers, but these short lists fail to cap-ture the contributions of many others.Perhaps the larger truth is that the pro-ject was its own reward. We had anopportunity to work with enormouslytalented people, on problems that wereboth challenging and important. I sus-pect that most of us remember theAMPS project as the best period in ourcareers, and look with justifiable prideat its profound results.

REFERENCES[1] J. S. Engel, “The Early History of Cellular

Telephone,” IEEE Commun. Mag., Aug. 2008,pp. 27–29.

[2] R. H. Frenkiel, Cellular Dreams and CordlessNightmares; Life at Bell Laboratories in Inter-esting Times, http://www.winlab.rutgers.edu/~frenkiel/dreams

[3] W. C. Jakes, Jr., Microwave Mobile Commu-nications, Wiley, 1974.

[4] High-Capacity Mobile Telephone SystemTechnical Report, AT&T FCC filing, Dec. 20,1971.

[5] V. H. MacDonald, “The Cellular Concept,”Bell Sys. Tech. J., vol. 58, no. 1, Jan. 1979.

[6] G. A. Arredondo, J. C. Feggeler, and J. I.Smith, “Voice and Data Transmission,” BellSys. Tech. J., vol. 58, no. 1, Jan. 1979.

[7] Z. C. Fleur and P. T. Porter, “Control Archi-tecture,” Bell Sys. Tech. J., vol. 58, no. 1,Jan. 1979.

[8] K. J. S. Chadha et al., “Mobile TelephoneSwitching Office,” Bell Sys. Tech. J., vol. 58,no. 1, Jan. 1979.

[9. N. Ehrlich, R. E. Fisher, and T. K. Wingard,“Cell-Site Hardware,” Bell Sys. Tech. J., vol.58, no. 1, Jan. 1979

[10. R. H. Frenkiel, “Cellular Radiotelephone Sys-tem Structured for Flexible Use of DifferentCell Sizes,” Patent 5,144,411; Mar. 13, 1979.

[11. D. L. Huff, “The Developmental System,”Bell Sys. Tech. J., vol. 58, no. 1, Jan. 1979.

[12. R. E. Fisher, “A Subscriber Set for the Equip-ment Test,” Bell Sys. Tech. J., vol. 58, no. 1,Jan. 1979.

[13. J. T. Walker, “A Service Test Mobile Tele-phone Control Unit,” Bell Sys. Tech. J., vol.58, no. 1, Jan. 1979.

[14. G. C. DiPiazza, A. Plitkins, and G. I. Zysman,“The Cellular Test Bed,” Bell Sys. Tech. J.,vol. 58, No. 1, Jan. 1979.

[15. Advanced Mobile Phone Service: Develop-mental System Reports; AT&T FCC filing,June 9, 1977, and each 90 days thereafter.

BIOGRAPHYRICHARD FRENKIEL [F] received his B.S. degree inmechanical engineering from Tufts University in1963 and his M.S. degree from Rutgers Universi-ty in 1965. At Bell Labs, his involvement in cellu-lar systems began in 1966 and lasted until1983. He was one of the authors of AT&T’s1971 cellular proposal to the FCC and the inven-tor of an efficient method for cell splitting dur-ing cellular growth. He became head of MobileSystems Engineering in 1977, and served on theEIA committee that proposed the nationwidecompatibility standard for cellular. In 1983 hebecame head of R&D for AT&T’s cordless tele-phone business unit, where he led the teamthat designed the very successful 5000 series ofcordless telephones, and he was responsible fortheir initial production in Singapore. Followinghis retirement from Bell Labs in 1993 he joinedthe Wireless Information Networks Laboratory atRutgers, where he still teaches a course in wire-less business strategy. In 1999 he served asMayor of Manalapan Township, New Jersey. Forhis work on cellular systems and cordless tele-phones, he has received a number of awards,including the National Medal of Technology(1994), the Alexander Graham Bell Medal of theIEEE (1987), and the Achievement Award of theIndustrial Research Institute (1992). He is amember of the National Academy of Engineer-ing.


24 IEEE Communications Magazine • September 2010



Documents

INTRODUCTION BY THE EDITOR - WINLABnarayan/Course/Wireless... · 2019-07-25 · cell phone manufacturers create the new radio designs these phones would require. This concession gained