Wideband Radar for Ballistic Missile Defense and Range-Doppler

VOLUME 12, NUMBER 2, 2000 LINCOLN LABORATORY JOURNAL 267

• CAMP, MAYHAN, AND O’DONNELLWideband Radar for Ballistic Missile Defense and Range-Doppler Imaging of Satellites

Wideband Radar for BallisticMissile Defense and Range-Doppler Imaging of SatellitesWilliam W. Camp, Joseph T. Mayhan, and Robert M. O’Donnell

■ Lincoln Laboratory led the nation in the development of high-powerwideband radar with a unique capability for resolving target scattering centersand producing three-dimensional images of individual targets. The Laboratoryfielded the first wideband radar, called ALCOR, in 1970 at Kwajalein Atoll.Since 1970 the Laboratory has developed and fielded several other widebandradars for use in ballistic-missile-defense research and space-objectidentification. In parallel with these radar systems, the Laboratory has developedhigh-capacity, high-speed signal and data processing techniques and algorithmsthat permit generation of target images and derivation of other target features innear real time. It has also pioneered new ways to realize improved resolution andscatterer-feature identification in wideband radars by the development andapplication of advanced signal processing techniques. Through the analysis ofdynamic target images and other wideband observables, we can acquireknowledge of target form, structure, materials, motion, mass distribution,identifying features, and function. Such capability is of great benefit in ballisticmissile decoy discrimination and in space-object identification.

T for thedevelopment of wideband radar systems wasrooted in the success of high-power instru-

mentation radars for research in ballistic missile de-fense (BMD) and satellite surveillance. In 1962 theTarget Resolution and Discrimination Experiment(TRADEX) radar, which was modeled in part afterthe Millstone Hill radar in Westford, Massachusetts,and built by RCA, became operational at KwajaleinAtoll in the Marshall Islands. This UHF and L-bandradar was the primary sensor for the Advance Re-search Projects Agency (ARPA)–sponsored PacificRange Electromagnetic Signature Studies (PRESS)project, which Lincoln Laboratory managed for theU.S. Army in support of its BMD program [1].

Project PRESS emphasized the use of radar andoptical sensors for the observation, tracking, measure-

ment, and characterization of a full-scale interconti-nental ballistic missile (ICBM) targets. A major effortof BMD research at Lincoln Laboratory was the de-velopment of techniques to discriminate warheadsfrom penetration aids, or penaids. These penaids,which accompany warheads in a ballistic missile reen-try complex, are devices designed to confuse, blind,overwhelm, or otherwise prevent defense systemsfrom identifying and destroying the warheads. TheLaboratory worked both on the development of pen-aids for U.S. missile systems and on techniques forreal-time discrimination of potential enemy penaidsthat could be used against U.S. defense systems. Amajor penetration aid at that time was the decoy, adevice that looked to the radar like a real warhead.

Decoys could be quite sophisticated and complex,but they all conformed to the basic requirement of


268 LINCOLN LABORATORY JOURNAL VOLUME 12, NUMBER 2, 2000

being light in weight compared to the weight of awarhead. Thus the fundamental approach to warheadidentification was to discriminate between warheadsand penaids on the basis of motion, size, and shapedifferences caused by these weight constraints. In theearly days these decoys were designed to mimic thedynamics and sensor signatures of warheads, and theywould challenge a defense radar’s ability to discrimi-nate between similar targets in reentry. A second ap-proach to warhead discrimination involved traffic de-coys, which consisted of a large number of smallerobjects designed to overwhelm and confuse a BMDsystem. To defeat traffic decoys, the radar needed todismiss a large number of less credible objects quickly.Wideband radar operation helped with both of thesediscrimination tasks [2].

By the mid-1960s the TRADEX radar had proveninvaluable as a sensor capable of highly accuratetracking of ballistic missile components from horizonbreak through midcourse and reentry. It could alsoidentify characteristic signatures of such componentsthrough processing of radar returns from reentry bod-ies and their accompanying ionized wakes. Duringthis same era many other field radars and laboratoryresearch facilities were gathering volumes of data onreal and simulated ballistic missile components andreentry wakes. As these data were analyzed, it becameclear that low-range-resolution radar systems operat-ing at the then-available frequencies were inadequateto unambiguously identify target signatures suitablefor discrimination in a BMD environment.

In particular, effective discrimination against stra-tegic threats requires a means of dealing with poten-tially large numbers of small decoys and penaids athigh altitudes well beyond the level where atmo-spheric deceleration becomes a discriminator betweenheavy and lightweight objects. This need was espe-cially acute in the 1960s, when BMD emphasis wason wide-area defense of cities. It was recognized thatdiscrimination radars with wide bandwidth and thecorresponding fine range resolution would be able tomeasure the lengths of objects, and quickly identifyand eliminate radar targets substantially smaller thanwarheads from consideration as threats. Furthermore,the high operating frequencies required for wide-bandwidth radars would be of added benefit over

those available in Nike Zeus and TRADEX (both atL-band). Bandwidths that are 10% of the radar’s car-rier frequency are reasonably straightforward toimplement (e.g., 500 MHz at C-band or 1000 MHzat X-band). At the higher frequencies required forwide-bandwidth sensing, there is greater potential tocharacterize the radar target’s physical features, thusproviding another level of capability for discoveringattempts to disguise a target’s true nature. In addition,higher operating frequencies are less vulnerable thanlower frequencies to the effects of nuclear blackout.

Given these considerations, Lincoln Laboratoryinitiated programs in the 1960s for the developmentof wideband high-resolution, high-power radars op-erating at high microwave frequencies. The first radarto be deployed was the ARPA-Lincoln C-bandObservables Radar, or ALCOR. It was designed as aninstrumentation radar to support research in BMDwideband discrimination techniques, and it achievedoperational status in 1970 at Kwajalein Atoll [3].

Another major thrust at Lincoln Laboratory in thelate 1950s and 1960s—an effort that eventually ledto specialized wideband radar systems—was in thearea of satellite surveillance. During this era thespace-defense establishment and NASA grew con-cerned about the plethora of satellites and debris or-biting the earth. Lincoln Laboratory and others rec-ognized that radar sensors beyond the capabilities ofexisting systems would be needed for space-objectidentification. The use of wideband techniques couldallow specific scattering centers to be identified.These techniques, coupled with the development ofhigh pulse-repetition-frequency coherent waveforms,would make the generation of detailed radar imagesof orbiting space objects possible.

The truth of this claim became abundantly clearshortly after ALCOR came on line. The space-sur-veillance community had arranged to enlist ALCORin tracking satellites on a noninterference basis. Thenin 1970 China launched its first satellite, which wasobserved by ALCOR. Analysis of ALCOR images ofthe booster rocket body revealed the dimensions ofthis object. This information was of great interest tothe Department of Defense because it gave insightinto the size and payload capacity of the forthcomingChinese ICBMs. This observation, which was a his-



toric first for the defense establishment, resulted insatellite imaging missions becoming an integral partof ALCOR operations.

More recently, the use of wideband phased-arrayradars has greatly facilitated the transition of BMDweapons systems from nuclear to non-nuclear, hit-to-kill interception techniques. These radars use modernsolid state microwave technology and high-capacityhigh-speed computers and signal processors, all ofwhich permit near-real-time imaging and discrimina-tion processing on a large number of targets.

Wideband Observables

The distinguishing characteristic of a wideband radaris its fine range resolution, which is inversely propor-tional to the operating bandwidth. Such a radar sys-tem has a range resolution that is a fraction of the lin-ear dimensions of its intended targets. These radarsgenerally operate at high frequencies, where wide-bandwidth waveforms are easier to implement. Withrange resolution fine enough to encompass a target ina significant number of resolution cells, it becomespossible to distinguish individual scattering centers,which occur at regions of physical discontinuity. Aballistic missile warhead, for example, exhibits radarreflections from the nose, body joints, and base, aswell as other points of discontinuity such as antennaports. To observe radar reflections from smaller dis-continuities, the radar must be able to operate atshort wavelengths, since discontinuities much smallerthan a wavelength will in general produce low-inten-sity reflected signals from the target. In addition, ashort wavelength is desirable for observing curvedsurfaces, because when the wavelength is short com-pared to the radius of curvature, the radar reflection isdominated by specular reflection, thus allowing afiner determination of the size and shape of corre-sponding surfaces.

For an object that reenters the atmosphere andgenerates an ionized wake, fine range resolution al-lows examination of the wake in thin slices, which re-sults in the separation of reflections from different at-mospheric phenomena around the hard body, such asthe plasma layer at the nose or leading surface, theplasma sheath around the body, the boundary layers,the shock fronts, and the development of turbulent

regions at the rear of the body. Such observations areof great value in deriving information about a target’sphysical parameters, structural and heat-shield mate-rials, and the function of reentering objects, all ofwhich aid the discrimination process of distinguish-ing warheads from decoys.

In the above scenario, the radar produces a one-di-mensional range profile of the target. However, if thetarget is rotating about an axis that has a componentperpendicular to the radar line of sight, such thatsome scattering centers are moving toward the radarwith respect to others that are moving away from it,we can construct a one-dimensional cross-range pro-file for each range cell through Doppler processing ofthe radar returns. The range and cross-range profilescan then be combined to produce a two-dimensionalrange-Doppler image of the complete body. We cananalyze this image to yield body size, body shape, theposition and nature of scattering centers, the presenceof internal reflections, the rotation rates, and the rota-tion axes for the object. In addition, these images canprovide valuable information on the nature of thematerials used in constructing the body, and informa-tion about antennas, apertures, and interior struc-tures of such an object.

Three-dimensional images can be generated fromthe two-dimensional images by using a techniquecalled extended coherent processing. With this tech-nique a series of range-Doppler images are collectedover a time period when the target presents differentlook angles to the radar. The series of range-Dopplerimages is then coherently processed and referenced toa particular look angle. The resulting three-dimen-sional images produce even greater detail of target fea-tures than the two-dimensional range-Doppler im-ages. The image of the damaged Skylab orbitinglaboratory shown in Figure 1 is an example of thiskind of processing.

More recent advances in signal processing hard-ware and computational speed have led to the genera-tion and measurement of wideband observables inreal time. These observables, which can be used forreal-time BMD discrimination, include determina-tion of body length, feature identification, and radarimages. Doppler processing and coherent phase-de-rived range techniques permit real-time indications of



macro- and micro-dynamic body motion, which mayoffer clues to mass and mass distribution. Advances inwideband phased-array radar design now make it pos-sible to exploit wideband observables on multiple ob-jects in a missile complex.

ALCOR

The initial wideband radar research at Lincoln Labo-ratory was embodied in a program called WidebandObservables. Initial objectives of this program were toverify the nature of wideband returns from realistictargets and investigate technology challenges thatmight arise in developing full-scale wideband radarsystems. To this end the Laboratory constructed aground-based static radar range that included a smallwideband radar and facilities for mounting and rotat-ing real, replica, and simulated targets. The promisingresults of this program and the pressing need for afine range-resolution instrumentation radar at Kwaja-lein led to the decision to develop the ARPA-LincolnC-band Observables Radar, or ALCOR.

ALCOR, shown in Figure 2, was the first high-power, long-range, wideband field radar system. Lin-coln Laboratory was the prime contractor forALCOR; a variety of industrial firms provided majorsubsystems and hardware (such as Hughes, Westing-house, Honeywell, and RCA). It became operationalat Kwajalein Atoll in 1970, and was probably the firstwideband radar in the world to reach that status, al-

though research in this field was under way elsewhereduring the 1960s, most notably at Rome Air Devel-opment Center. ALCOR was designed to have aslarge a bandwidth, sensitivity, range coverage, andtracking agility as contemporary technology wouldreasonably allow, to support its role in BMD research.It was located next to the TRADEX radar on Roi-Namur Island in the Kwajalein Atoll. Figure 2(a)shows the sixty-eight-foot-diameter ALCOR radome,and Figure 2(b) shows the forty-foot ALCOR an-tenna and its pedestal inside the radome.

ALCOR operates at C-band (5672 MHz) with asignal bandwidth of 512 MHz that yields a rangeresolution of 0.5 m. (The ALCOR signal was heavilyweighted to produce low range sidelobes with theconcurrent broadening of the resolution.) Its wide-bandwidth waveform is a 10-µsec pulse linearly sweptover the 512-MHz frequency range. High signal-to-noise ratio of 23 dB per pulse on a one-square-metertarget at a range of a thousand kilometers is achievedwith a high-power transmitter (3 MW peak and 6kW average) and a forty-foot-diameter antenna.Cross-range resolution comparable to range resolu-tion is achievable with Doppler processing for targetsrotating at least 3° in the observation time. The pulse-repetition frequency of this waveform is two hundredpulses per second.

Processing 500-MHz-bandwidth signals in someconventional pulse-compression scheme was not fea-sible with the technology available at the time ofALCOR’s inception. Consequently, it was necessaryto greatly reduce signal bandwidth while preservingrange resolution. This is accomplished in a time-bandwidth exchange technique (originated at the Air-borne Instrument Laboratory, in Mineola, New York)called stretch processing [4], which retains range reso-lution but restricts range coverage to a narrow thirty-meter window. In order to acquire and track targetsand designate desired targets to the thirty-meterwideband window, ALCOR has a narrowband wave-form with a duration of 10.2 µsec and bandwidth of6 MHz. This narrowband waveform has a muchlarger 2.5-km range data window.

The ALCOR beamwidth is 5.2 milliradians, or0.3°. This beamwidth, together with a high-perfor-mance antenna mount, enables ALCOR to produce

FIGURE 1. Simulated radar image (actual radar images ofsatellites remain classified) of the NASA Skylab orbitinglaboratory, with a damaged solar panel on one side and apartially deployed solar panel on the other.



precision target trajectories and provide high-qualitydesignation data to the other Kwajalein radars. Thisvery narrow beam also caused some real challenges intarget search and acquisition. Searching with a 5.2-milliradians beam is akin to looking through a two-hundred-foot-long pipe that is only one foot in diam-

eter. Fortunately, ALCOR was located next toTRADEX and ALTAIR, which are much bigger ra-dars with great search capability.

The more difficult technical challenges includedconstruction of the analog 10-µsec, 512-MHz-band-width linear-FM-ramp generator. It required a highdegree of timing stability, phase coherence, and fre-quency linearity to control range sidelobes to a levelof at least 35 dB below the peak response. The timinggenerator for controlling ramp triggering likewise re-quired a high degree of precision and stability to limittarget tracking jitter to a small fraction of a range-resolution cell.

Also, phase distortion in the transmitter and mi-crowave systems is held to a low level through carefuldesign and matching of components, in order to real-ize low range sidelobes. Residual distortion is com-pensated by transversal equalizers in the wideband re-ceiver channels. Another feature that is part of theALCOR radar design and unique to wideband sys-tems is compensation for waveguide dispersion.

ALCOR is one of the earliest radars to incorporatecomputers running real-time programs as an integralpart of radar operations. Major radar control and tim-ing functions are under computer control, and therange and angle-tracking loops are closed through thecomputer. In recent years advances in signal process-ing technology applied to ALCOR have made it pos-sible to generate wideband images of multiple targetsin multiple range windows in near real time. [3, 5]

TRADEX S-Band

In 1972 the TRADEX UHF radar was replaced by anS-band system (built by RCA under the direction ofLincoln Laboratory) while retaining its L-band capa-bility. Although TRADEX S-band was the secondwideband radar system to be brought on line by Lin-coln Laboratory, the approach taken to achieve finerange resolution was substantially different from thattaken in ALCOR and later wideband systems. The S-band wideband waveform grew out of research on fre-quency-jumped pulses conducted by the Laboratoryin the 1960s. The new set of S-band waveforms in-cludes one with a signal bandwidth of 250 MHz.This bandwidth is achieved by transmitting a string,or burst, of 3-µsec pulses; each pulse has a different

FIGURE 2. (a) The sixty-eight-foot-diameter ALCOR radomeand (b) the forty-foot-diameter ALCOR antenna and its ped-estal. ALCOR, which became operational at Kwajalein Atollin 1970, was the first high-power, long-range, wideband fieldradar.

(a)

(b)



center frequency such that the bandwidth spanned bythe burst is 250 MHz. The spacing and number ofpulses in a burst is variable and the maximum repeti-tion rate is one hundred bursts per second. Coherentintegration of a burst yields a range resolution of onemeter. Processing of this frequency-jumped-bursttype waveform at S-band is implemented post-mis-sion. Wideband target profiles have been constructedpost-mission and have helped demonstrate the utilityof wideband observables for use as a target discrimi-nant at S-band.

The Haystack Long-Range Imaging Radar

From its earliest days of operation ALCOR was calledupon to track and image satellites, both domestic andforeign. ALCOR’s historic imaging of the boosterrocket body of China’s first orbiting satellite in 1970was followed by similar success in imaging theUSSR’s Salyut-1 space station in 1971. The demon-stration of such imaging capability led to the estab-lishment of the Space Object Identification programat Lincoln Laboratory. One of the many successes ofthis program was the imaging of NASA’s troubledSkylab orbiting laboratory shortly after its launch in1973, as shown in Figure 1. Telemetry data fromSkylab showed that something was wrong with thedeployment of the solar panels. ALCOR imagesshowed that one solar panel was missing and theother panel was only partially deployed, and there wasno evidence of the presence of the micrometeoriteshield. This information was extremely useful toNASA in determining how to recover from this mis-hap and successfully continue the Skylab mission.

As the Space Object Identification program con-tinued, the capability to image satellites out to deep-space ranges soon became an important requirement.The limited sensitivity of ALCOR allowed the obser-vation of satellites only out to intermediate altitudes.At the same time, researchers wanted to improverange resolution, extend range-window coverage, andachieve higher pulse-repetition frequencies in orderto eliminate cross-range ambiguities in the images ofrapidly rotating space objects. Concurrent advancesin signal processing technology gave promise thatsuch improvements in radar techniques could readilybe accommodated. After exploring a number of op-

tions, researchers determined that a high-perfor-mance long-range radar imaging capability could beprovided by a new radar system added to Haystack,the Lincoln Laboratory–built radio-astronomy, com-munication, and radar research facility located inTyngsboro, Massachusetts. The development of thisnew radar capability for Haystack was sponsored byARPA. After the facility was completed in 1978, op-erations were supported by the U.S. Air Force.

The Haystack system has a number of features thatrendered this option extremely attractive. It has alarge diameter (120 ft) antenna needed to achievedeep-space ranges. The antenna was designed withCassegrainian optics and could accommodate plug-inradio-frequency (RF) boxes at the vertex of the pa-raboloidal dish. These boxes supported various com-munications, radio astronomy, and radar functions.The interchangeable boxes are 8 × 8 × 12 ft, which islarge enough for the high-power (400 kW peak and200 kW average) new transmitter and associated mi-crowave plumbing, feedhorns, and low-noise receiv-ers needed for the long-range imaging radar.1 TheHaystack antenna surface tolerance allows efficientoperation up to 50 GHz, thus readily supporting op-erating at X-band (10 GHz) with a bandwidth of1024 MHz, and a resulting range resolution of 0.25m. A system for interchanging ground-based elec-tronics and power sources supporting the various RFboxes was already in place. Using an established facil-ity with existing antenna and prime power sourcesgreatly reduced the cost of the new system, known asthe Long Range Imaging Radar, or LRIR [6].

The LRIR, which was completed in 1978, is ca-pable of detecting, tracking, and imaging satellitesout to synchronous-orbit altitudes, approximately40,000 km. The range resolution of 0.25 m ismatched by a cross-range resolution of 0.25 m for tar-gets that rotate at least 3.44° during the Doppler-pro-cessing interval. The wideband waveform is 256 µsec

1 Over the years several modifications have been made to theHaystack transmitter to make it more reliable. These haveaffected the radar’s average power and maximum pulse width.Currently, the maximum average power of the radar is 140 kWand the maximum pulse width is 5 msec. With these values theradar continues to perform all of its required missions.



long and the bandwidth of 1024 MHz is generated bylinear frequency modulation. The pulse-repetitionfrequency is 1200 pulses per second. The LRIR em-ploys a time-bandwidth exchange process similar tothat of ALCOR to reduce signal bandwidth from1024 MHz to a maximum of 4 MHz, correspondingto a range window of 120 m, while preserving therange resolution of 0.25 m. To place a target in thewideband window, we first acquire the target with acontinuous-wave acquisition pulse that is variable inlength from 256 µsec (for short-range targets) to 50msec (for long-range targets). An acquired target isthen placed in active tracking by using 10-MHz-bandwidth chirped pulses, again of variable length,from 256 µsec to 50 msec. The wideband window isthen designated to the target. Antenna beamwidth is0.05°. Figure 3(a) shows an artist’s rendition of the120-foot Haystack antenna in its 150-foot radome;Figure 3(b) shows a photograph of the LRIR feedhorn and transmitter/receiver RF box in the Haystackradome.

The LRIR design and construction at Haystack re-quired special attention to waveform distortion,phase control, phase stability, timing precision, andtiming stability (as did ALCOR). With the longerwideband pulse width and advances in digital signalprocessor and computer technology, however, muchof the signal processing and compensation has beenaccomplished with digital hardware under computercontrol.

The transmitter and microwave subsystems alsopresented design challenges. To achieve maximumsensitivity the transmitter used four traveling-wavetubes (TWT) operating in parallel. Each TWT high-power amplifier developed by Varian for LRIR gener-ates 100 kW peak and 50 kW average X-band power.Combining and balancing the TWT outputs throughthe myriad of required microwave components whilecontrolling phase errors was a major challenge.

The front-end receiver amplifiers developed byAirborne Instrument Laboratory are cryogenicallycooled parametric amplifiers, or paramps. These effi-cient paramps are major contributors to Haystack’shigh radar sensitivity, achieving a system noise tem-perature of 35 K.

Although not a directly related part of the researchand development of LRIR, the Haystack antenna andits protective radome are impressive engineering ac-complishments. The Haystack 150-ft-diameter ra-dome, at the time it was built in the early 1960s, wasthe largest rigid radome in the world. It was designedby the ESSCO Company of Concord, Massachusetts,to survive 130-mph winds.

The Haystack antenna was also considered an en-gineering feat at the time of its construction. It wasthe first large structure designed by computer with afinite-element representation of every strut in itsstructure. North American Aviation and the MITCivil Engineering Department accomplished this sig-nificant first.

FIGURE 3. (a) Artist’s rendition of the 120-foot Haystack antenna in its 150-foot radome; (b) the long-range imaging radar(LRIR) feed horn and transmitter/receiver radio-frequency (RF) box in its test dock in the Haystack radome.

(a) (b)



The Haystack Auxiliary Radar

When the LRIR became operational in 1978, theHaystack facility was being operated by a universityconsortium known as the Northeast Radio Observa-tory Corporation, or NEROC. The primary missionof Haystack was radio astronomy. Lincoln Laboratorycontracted with NEROC to use the facility for satel-lite tracking and imaging for a thousand hours peryear. Eventually, U.S. Space Command and NASArecognized that this arrangement for sharing Hay-stack was too restrictive to satisfy their needs. Therestriction was especially limiting when there wasan immediate need to assess new or unexpected for-eign launches, or when space debris needed to becatalogued.

The Haystack Auxiliary Radar (HAX) system,shown in Figure 4, was conceived to eliminate this re-striction on observation time, and simultaneouslyfurther improve range resolution and pulse-repetitionfrequency. HAX operates at Ku-band with its ownforty-foot antenna, transmitter, RF hardware, and re-ceiver, but it shares the LRIR control and signal anddata processing systems with Haystack. HAX, whichbegan operation in 1993, is the first radar to have asignal bandwidth of 2000 MHz, which improves therange resolution to 0.12 m. HAX represents a signifi-

cant advance in radar imaging capability, producingfiner and sharper images of satellites than the Hay-stack LRIR. It is also extremely useful in producingdetailed information for NASA on the locations, or-bits, and characteristics of space debris.

The Millimeter Wave Radar

The Millimeter Wave Radar, or MMW, was built atKwajalein by Lincoln Laboratory (with significantcontributions by the University of Massachusetts,RCA, and Raytheon) to extend the general imagingand tracking capabilities of ALCOR and to developmillimeter-wavelength signatures of ballistic missilecomponents. The MMW, shown in Figure 5, becameoperational at Ka-band (35 GHz) in 1983, and W-band (95.48 GHz) in 1985, sharing a paraboloidalantenna with a diameter of forty-five feet. Both sys-tems initially featured wideband waveforms of 1000-MHz spread generated by linear FM, and achieved0.28-m range resolution. The transmitted pulsewidth is 50 µsec at a maximum pulse-repetition rateof 2000 pulses per second. The initial peak power atKa-band was 60 kW and at W-band was 1.6 kW.

A major thrust in the evolution of the MMW ra-dar has been to demonstrate the feasibility of candi-date real-time discrimination algorithms required forfire control and guidance of hit-to-kill BMD inter-ceptors. To this end, the radar was designed with arigid mount and narrow beam to provide preciseangle metric accuracy (≤50 µradians). Several con-tractors assisted the Laboratory in the development ofthe MMW radar, among them researchers at the Uni-versity of Massachusetts, RCA (now Lockheed Mar-tin), and Raytheon. The combination of metric accu-racy, wide bandwidth, and high Doppler-resolutioncapabilities makes MMW an excellent sensor for areal-time discrimination test bed. It provides ex-tremely accurate estimates of motion differencescaused by mass imbalances on real and threat-like tar-gets and other feature-identification processing. Sucha real-time test bed, called the Kwajalein Discrimina-tion System, was implemented and exercised atMMW from 1988 through 1992. This system hasdemonstrated the feasibility of numerous real-timewideband discrimination algorithms. Following thissuccessful demonstration, many of these processing

FIGURE 4. The Haystack radar site in Tyngsboro, Massachu-setts. The 150-foot LRIR Haystack radome and 120-foot an-tenna are on the left, and the fifty-foot Haystack Auxiliary Ra-dar (HAX) radome, forty-foot HAX antenna, and equipmentbuilding are on the right.



algorithms have been subsequently implemented per-manently into the MMW real-time system.

Beginning in the late 1980s, a significant effortwas carried out to further enhance the capabilities ofthe MMW radar. Advances in computer technologyreached the point where real-time pulse compressionat high pulse-repetition frequencies was possible. Thiscapability results in improved sensitivity realized fromreal-time coherent and noncoherent pulse integrationat 35 GHz. In 1989 researchers implemented a digitalpulse-compression system that compresses everypulse in real time and subsequently improves coher-ent processing at higher tracking rates. A state-of-the-art beam-waveguide antenna feed replaced the morelossy conventional microwave-waveguide plumbing.A second 35-GHz tube was also added, whichdoubled the average transmitted power. These modi-fications increased the signal pulse detection range ona one-square-meter target to over two thousand kilo-meters. System bandwidth was also increased to 2GHz, resulting in a range resolution of about 0.10 m.

The 95-GHz MMW system is now undergoing amajor upgrade. Recent advances in MMW solid statetechnology, combined with state-of-the-art quasiopti-cal feed elements, have resulted in significant in-creases in system sensitivity. Mixer diodes capable of

cryogenic operation have led to a reduction in the re-ceiver noise figure, and a Gunn-effect diode interme-diate-power amplifier has boosted transmit power bydriving the final TWT into saturation. Application ofbeam-waveguide optics techniques to the antenna hasresulted in a reduction in transmit and receive losses,while simultaneously increasing transmit and receiveisolation and power-handling capability. In all, theseimprovements provide a sensitivity improvement ofalmost two orders of magnitude, yielding the highercapability for metric tracking and range-Doppler im-aging on a variety of important targets. When thesemodifications are completed, the robust combinationof 35-GHz and 95-GHz wideband capability on asingle sensor will ensure MMW’s place as one of theworld’s premier wideband-imaging systems [7, 8].

Cobra Dane and Cobra Judy

Two of the most important sensors in the evolution ofwideband BMD architecture and technology are theCobra Dane and Cobra Judy wideband phased-arrayradars. Both of these radars were built by theRaytheon Corporation under support of the U.S. AirForce and U.S. Army. Lincoln Laboratory played asupporting role in the development of these phased-array sensors, and it continues to provide essentialsupport in upgrades and data processing. The Labora-tory has also been heavily involved in the analysis ofdata. Developed to collect data on strategic ballisticmissiles, these sensors were at the forefront ofwideband phased-array technology. Cobra Dane,shown in Figure 6, was the first sensor to become op-erational, in 1976. Located on the island of Shemya,Alaska, in the western end of the Aleutian chain, Co-bra Dane can observe the exo-atmospheric portionsof Russian missile flights into the Kamchatka Penin-sula and the Pacific Ocean.

The Cobra Judy mobile S-band phased-array ra-dar, located on the U.S. Naval ship ObservationIsland, was subsequently developed and became op-erational in 1981 to provide improved global ballisticmissile data collection on a mobile sea-based plat-form. An X-band wideband dish radar added to Co-bra Judy became operational in 1984. The shift to X-band was to enhance resolution and imagingcapability relative to S-band, and subsequently to

FIGURE 5. The forty-five-foot millimeter-wave (MMW) an-tenna and its sixty-eight-foot radome under construction onRoi-Namur Island at Kwajalein. The MMW radar was built toextend the general imaging and tracking capabilities ofALCOR and to develop millimeter-wavelength signatures ofballistic missile components.



support the BMD community in developing a na-tional missile defense system that would incorporatean X-band phased array for interceptor fire control.Cobra Judy, shown in Figure 7, remains to this datethe world’s largest mobile phased-array sensor.

Lincoln Laboratory’s support of the developmentand use of these high-quality instrumentation radarsinvolved a number of significant contributions. Spe-cifically, Lincoln Laboratory (a) developed the re-quirements and specifications for each of the sensorsand validated sensor performance after deployment;

(b) developed the data processing and data distribu-tion architecture, including calibrations and the dataprocessing software and methodology; (c) developedand applied key analysis tools tailored for BMD dataexploitation; and (d) continues to participate in theoperational use of these sensors. Some of the analysistools tailored for BMD data exploitation are syntheticbandwidth expansion for enhanced range resolution,extended coherent processing for three-dimensionalimaging, macro- and micro-motion-dynamics pro-cessing techniques, and enhanced spectral estimationfor feature identification. The Laboratory continuesto participate in the operational use of these sensorsthrough Cobra Judy mission planning, ballistic mis-sile profile development and operator training, CobraJudy upgrades and enhancements, and the develop-ment of data interpretation and analysis techniques.

Development of the Cobra Dane and Cobra Judysensors required advances in two key technology ar-eas: wideband wide-angle phased-array scanning cov-erage, and automated target classification and datacollection.

Cobra Gemini

Over the past few years, many nations have begun toobtain tactical ballistic missiles. The recent Gulf Warillustrated that these missiles can be important politi-cal weapons when equipped with explosive warheads.

FIGURE 7. The wideband phased-array Cobra Judy radar on the U.S. Naval ship Observation Island. TheS-band phased-array radar is at the stern of the ship; the mechanically scanned dish just forward of it isthe X-band radar antenna. The Cobra Judy is the world’s largest mobile phased-array sensor.

FIGURE 6. The wideband phased-array Cobra Dane radar,located on the island of Shemya, Alaska, in the western endof the Aleutian chain. In this location, Cobra Dane can ob-serve the exo-atmospheric portions of Russian missileflights into the Kamchatka Peninsula and the Pacific Ocean.



If these missiles are coupled with a chemical, bacte-riological, or nuclear weapons capability, they presentan extremely threatening weapons system. Currentdevelopment of theater-missile defense systems suchas Theater High Altitude Area Defense requires dataon these tactical missile systems for use in fire-con-trol-radar discrimination functions.

In 1996, to meet the defense community’s globalrequirements for radar signature data for theater-mis-sile defense, Lincoln Laboratory, with sponsorshipfrom the U.S. Air Force Electronic Systems Center,initiated development of an operational prototype ofa transportable instrumentation radar called CobraGemini. Cobra Gemini is a sensor designed to collectmetric and signature data on tactical theater-ballistic-missile targets. Since an operational radar must re-spond to a wide variety of launch locations, a sensorlike the Cobra Gemini must be both air- and ground-transportable. It must also be capable of worldwideoperation on land and deck-mounted operation onships at sea. Figure 8 shows the Cobra Gemini radarimplemented in a sea-based configuration on the In-vincible, a U.S. Naval T-AGOS ship.

The signature data collection requirements includewideband and narrowband radar data at both S-bandand X-band. High pulse-repetition-rate, wide-band-width data are collected at both X-band and S-bandfrequencies to support detailed analysis of tactical

missile dynamics as well as the characterization of ob-jects in the missile threat complex.

The Cobra Gemini radar was designed to acquire a–5-dBsm (decibels relative to one square meter) targetat a range of a thousand kilometers over a 20°-widehorizon-surveillance fence. The thousand-kilometerrequirement for detection range is adequate for de-tecting tactical ballistic missiles over a wide range ofscenarios. Surveillance and target acquisition are per-formed at S-band by using the radar’s mechanicalscanning dish antenna, which has a fifteen-foot aper-ture. The radar has 50 kW of average power at S-bandand 35 kW at X-band. The antenna uses a dual-bandfeed to transmit both S-band and X-band energyfrom the same dish. After acquisition, tracked objectsare classified in real time so that high-resolution datacan be collected on high-priority objects. In thewideband mode, the radar has a bandwidth of 1 GHzat X-band and 300 MHz at S-band. These band-widths correspond to range resolutions of approxi-mately 0.25 m and 0.80 m, respectively.

The purpose of the Cobra Gemini program was toproduce a prototype radar with a short deliveryschedule and low risk of delays in development. Addi-tional radars of this type can then be more confi-dently produced by industry with this baseline de-sign. The Cobra Gemini prototype completed itsground-based test and evaluation in July 1998. The

FIGURE 8. The Cobra Gemini radar (large radome) in its sea-based configuration on the Invincible, a U.S.Naval T-AGOS ship. A transportable sensor such as Cobra Gemini is designed to assist in the detectionand discrimination of tactical theater ballistic missiles.



radar was then installed on the Invincible. The proto-type achieved operational capability on 1 March1999 after successful completion of its sea-based testand evaluation.

Advances in Wideband Data Analysis

Lincoln Laboratory has played a major role in the de-velopment of data-analysis techniques for use in ex-ploiting data collected by wideband radars on strate-gic and theatre ballistic missiles. The evolution incomplexity of discrimination technology has directlyfollowed the evolution of capability in computer pro-cessing. Figure 9 illustrates this evolution of analysistechniques over time and their relationship to the de-velopment of discrimination algorithms. As wide-band radar technology has evolved, providing higherpulse-repetition frequencies, wider bandwidth, andhigher RF operating frequencies, the analyst’s abilityto infer more information about a target under inter-rogation by the radar has increased significantly. De-velopments in computer processing have led to subse-quent real-time implementation of algorithms basedon these analysis techniques.

More recently, Lincoln Laboratory has developedand exploited several techniques for improving the

resolution of wideband coherent radar data. The firsttechnique uses modern spectral-analysis methods forimproving resolution relative to the restrictions ofconventional Fourier processing. These spectralmethods extrapolate signals in a radar-frequency di-mension by a process called bandwidth extrapolation.Each wideband pulse return includes the target fre-quency response over the chirped bandwidth. Mod-ern spectral-estimation techniques are then applied toextend this frequency response synthetically outsidethis band to a factor ranging from two to three timesthe bandwidth. This expanded pulse return is then re-compressed to provide finer range resolution (forpractical signal-to-noise ratios, an improvement of afactor of two to three in resolution is generally real-ized), and when applied to radar imaging, it providesmuch improved sharpness in the radar image [9].

The second technique uses signal processing mod-els that correspond to rotating-point motion. Themodels allow extended coherent processing over widetarget-rotation angles, resulting in improved Doppler(cross-range) resolution [10]. For sufficiently largeresolution angles and for constant-amplitude scatter-ing centers, extended coherent processing also im-proves the range resolution. Extended coherent pro-

FIGURE 9. The evolution of wideband radar data-analysis techniques and their relationship to the development of ballistic-mis-sile-defense discrimination algorithms. As wideband radar technology has evolved, providing higher pulse-repetition frequen-cies, wider bandwidths, and higher operating frequencies, the analyst’s ability to infer more information about a target underobservation by the radar has increased significantly.

1960s ~1970 ~1975 ~1983 ~1990 ~1997

Narrowband

RCS classifier(peak, mean)NarrowbandBallistic coefficient

Microdynamics(wobble)

Size/shapeWideband algorithm Length RCS Polarization Ballistic coefficient

MacrodynamicsBallistic coefficient usingphase-derived range

Wideband(low pulse- repetition

frequency)

Range-Dopplerimaging

Phase-derivedrange

Bandwidthexpansion

Ultra-wideband TimeAnalysis

techniques

Discriminationalgorithms

Time



cessing essentially aligns and stores radar pulses ob-tained over longer time spans as compared to conven-tional imaging. When combined with bandwidth ex-trapolation, extended coherent processing can achieveenhanced resolution in both range and Doppler(cross-range) spaces. For targets where the radar view-ing angle is at a constant aspect angle to the target’sangular-momentum vector, extended coherent pro-cessing provides high-quality three-dimensional radarimages.

More recently, the Laboratory has explored thepossibility of achieving ultrawideband resolution byusing data only over sparse subbands of the full ultra-wide bandwidth. We can view this technique as a gen-eralization of bandwidth extrapolation to multiplebands [10]. Ultrawideband’s potential as a discrimi-nation tool is much more robust, as scatterer-featureidentification on a specific target is inherently moreaccurate when observed over a much wider band-width.

Summary

Lincoln Laboratory has played a pioneering anddominant role in the development of high-powerwideband radar systems and technology. As the capa-bility of wideband radars has progressed, the Labora-tory has continued to develop high-speed signal pro-cessing devices and systems, and more sophisticatedanalysis techniques. These new developments permitthe formation of dynamic three-dimensional imagesof multiple targets in near real time, and have resultedin the development of discrimination algorithms ap-plied to real-time ballistic missile defense.

As a result of the Laboratory’s research and devel-opment efforts in wideband radar, signal processors,and data analysis, the original promise of widebandsystems has been fulfilled to an extent unforeseenthirty years ago. Today’s wideband systems reveal in-formation across the full spectrum of physical at-tributes of observed targets. The true test of the valueof this Lincoln Laboratory development is thattoday’s ballistic-missile-defense radars, such as theTheater High Altitude Area Defense radar and theNational Missile Defense Ground-Based Radar Pro-totype, feature substantial wideband capabilities thatare critical to meeting their challenging missions.

R E F E R E N C E S1. M.L. Stone and G.P. Banner, “Radars for the Detection and

Tracking of Ballistic Missiles, Satellites, and Planets,” Linc.Lab. J., in this issue.

2. P.A. Ingwersen and W.Z. Lemnios, “Radars for Ballistic MissileDefense Research,” Linc. Lab. J., in this issue.

3. M. Axelbank, W.W. Camp, V.L. Lynn, and J. Margolin,“ALCOR—A High-Sensitivity Radar with One-Half-MeterRange Resolution,” IEEE 71 International Conv. Dig., NewYork, 22–25 Mar. 1971, pp. 112–113.

4. M.I. Skolnik, Introduction to Radar Systems, 2nd ed. (McGraw-Hill, New York, 1980), p. 432.

5. R.K. Avent, J.D. Shelton, and P. Brown, “The ALCOR C-Band Imaging Radar,” IEEE Antennas Propag. Mag. 38 (3),1996, pp. 16–27.

6. E.C. Freeman, ed., MIT Lincoln Laboratory: Technology in theNational Interest (Lincoln Laboratory, Lexington, Mass.,1995), pp. 118–119.

7. M.D. Abouzahra and R.K. Avent, “The 100-kW Millimeter-Wave Radar at the Kwajalein Atoll,” IEEE Antennas Propag.Mag. 36 (2), 1994, pp. 7–19.

8. J.C. McHarg, M.D. Abouzahra, and R.F. Lucey, “95-GhzMillimeter Wave Radar,” SPIE 2842, 1996, pp. 494–500.

9. K.M. Cuomo, “A Bandwidth Extrapolation Technique forImproved Range Resolution of Coherent Radar Data,” ProjectReport CJP-60, Rev. 1, Lincoln Laboratory (4 Dec. 1992),DTIC #ADA-258462.

10. K.M. Cuomo, J.E. Piou, and J.T. Mayhan, “Ultra-WidebandCoherent Processing,” Linc. Lab. J. 10 (2), 1997, pp. 203–222.



. ’is a senior staff member in theSensor Systems and Measure-ment group. His recent re-search has involved the designof the Cobra Gemini radarand the modernization andremoting of the range instru-mentation radars at theKwajalein Missile Range. Hereceived an S.B. degree inphysics from MIT in 1963,and M.S. and Ph.D. degrees inphysics from the University ofPennsylvania in 1964 and1970. Before joining LincolnLaboratory in 1973, he was amember of the technical staffof the Radar Technologydepartment at the MITRECorp. In 1983, he left LincolnLaboratory to become man-ager of Radar Systems Engi-neering at RCA (nowLockheed Martin) inMoorestown, New Jersey. In1986, he joined the Interna-tional Signal and ControlGroup in Lancaster, Pennsyl-vania, as vice president, chiefscientist, where he was respon-sible for corporate research anddevelopment and long-rangetechnical planning for thegroup. In 1989 he returned toLincoln Laboratory. He is asenior member of the IEEEand is a present member andpast chairman of the IEEEAerospace and ElectronicSystems Society, Radar Sys-tems Panel, and is a memberof the Board of Governors ofthe IEEE Aerospace andElectronic Systems Society.His interests include classicalmusic and amateur radio.

. is the former leader of theRadar Technology group andassistant head of the RadarMeasurements division. Hereceived a B.S.E.E. degreefrom Virginia PolytechnicInstitute, and joined LincolnLaboratory in 1957. Afterinitial assignments in thedevelopment of radar receivercircuitry and radar beacons, hebecame active in radar systemsanalysis and design, contribut-ing to early concepts of ballis-tic missile defense (BMD) andsatellite surveillance by radar.He was a member of the teamthat evaluated the BallisticMissile Early Warning SystemSite I acceptance tests inThule, Greenland, and amember of the design teamthat established performancerequirements for the firstsatellite surveillance radar. AtKwajalein in the early 1960she was responsible for thehigh-power UHF/L-bandTRADEX radar. He wasappointed associate leader ofthe Range Measurementsgroup in 1964 and leader ofthe Radar Technology group in1969. During these years hisgroup managed the implemen-tation of ALTAIR at Kwajaleinand an extensive modificationof the TRADEX radar. In thelate 1960s he was projectleader for the ALCORwideband radar at Kwajalein,and in the mid-1970s hedirected the development ofthe Long Range ImagingRadar, used for imaging satel-lites out to synchronous-orbitaltitude. He retired fromLincoln Laboratory in 1992.

. is a senior member of thetechnical staff, and formerleader of the Sensor Systemsand Measurements group. Hereceived a B.S. degree fromPurdue University, and M.S.and Ph.D. degrees from TheOhio State University, all inelectrical engineering. In 1969,he joined the University ofAkron as an associate profes-sor. Since joining LincolnLaboratory in 1973, he hasworked in the areas of satellitecommunications antennas,adaptive antenna design andperformance evaluation, spa-tial spectral estimation usingmultiple-beam antennas, radarsystem design, radar dataanalysis, and electromagneticscattering from actively loadedtargets. He played a major rolein the design of the CobraGemini radar. At theLaboratory’s Kwajalein fieldsite, he served one four-yeartour as leader of the ALTAIRdeep-space tracking radar anda second tour as site managerof the Lincoln Laboratoryprogram. He has publishedextensively in the areas ofnonlinear interactions ofelectromagnetic waves inplasma media, adaptive an-tenna design, spectral estima-tion techniques, and electro-magnetic scattering. He twicereceived the R.W.P. KingAward from the IEEE Transac-tions on Antennas and Propa-gation Society (AP-S) forpapers published in the IEEETransactions on Antennas andPropagation, and has previ-ously served as a distinguishedlecturer for AP-S.

Documents

Wideband Radar for Ballistic Missile Defense and Range-Doppler