A hardware-compressive fiber-optic true time-delay steering system for phased-array antennas.
Phased-array antennas (PAA) consisting of elemental transmit/receive (T/R) modules offer many advantages, including steering without physical movement, accurate beam pointing and increased scan flexibility in two dimensions, reduced power consumption and weight, and precise elemental phase and amplitude control for maintaining low spatial sidelobes. These advantages have been recognized in radar and communications for some time, and during the last decade, a variety of PAA-based radar and communications systems have been successfully deployed worldwide. These systems operate in different frequency bands over the ELF (30 kHz) to EHF (60 GHz) ranges. Virtually all of these systems have used a phase shifter in each T/R module for controlling the phase of the radiated waveform and thereby steering the PAA.
Unfortunately phase shifters by their nature are accurate for any given phase angle at only one frequency, usually the PAA center frequency (design frequency). For other frequencies, the PAA pattern squints, that is, the beam peak angle is reduced for frequencies above the center frequency and increased for frequencies below the center frequency. In effect, this behavior restricts the instantaneous bandwidth (BW) of the PAA to much less than an octave, typically less than 10 percent of the PAA center frequency. In general, the PAA BW becomes smaller as the PAA becomes larger or as the scan angle increases.
The limited BW possible with phase shifters creates a major problem for next generation, high performance PAA-based radars, communications systems and multifunction front-end systems, which combine radar, communications and other functions in a common PAA. Such systems will be required to have large scan angles, for example [+ or -]70 [degrees], wide instantaneous BWs (100s of MHz to several GHz), center frequencies from UHF to X-band and the capability for multiple simultaneous, independent beams. In such systems, the actual number of T/R modules will depend on the system mission, as well as on the operating frequency, and is expected to be in the [10.sup.2] to [10.sup.4] range for all airborne, space-based, ground-based and shipboard systems.
To satisfy the wide BW requirements of such systems, true time delay (TTD) frequency-independent steering techniques must be used so that efficient vector summation in the receive mode or distribution in the transmit mode can be obtained independent of frequency. Fiber is an excellent medium for both TTD and signal distribution because it can store large BW analog signals ([is approximately]100 GHz) for long times (tens of [[micro]seconds]) and has low attenuation ([is less than] 0.6 dB/km at 1300 nm), which is flat over radio frequencies up to 100 GHz. In addition, fiber has good transmission stability by virtue of the small ratio of signal BW to optical carrier frequency; allows the remote processing of PAA signals, as well as optical wavelength division multi-plexing (WDM) to minimize the number of lines in the PAA feed link; is a nonconducting dielectric, and therefore, does not disturb the RF field; is secure and EMI immune; and is flexible. Fiber is also good because of low mass (0.07 g/m vs. 40 g/m for coaxial cable) and small volume, which are important attributes for airborne- and space-based systems.
The importance of the fiber-optic (FO) TTD has been widely recognized in both radar and communications. Independent demonstrations have already verified the viability of the FO TTD concept using both laboratory equipment(1,2) and an actual PAA.(3) Several developmental TTD systems have been implemented or are currently underway worldwide.(4-16) In developing FO TTD systems, much attention must be given to the amount of hardware. For large PAAs, straightforward FO TTD implementations result in unrealistically large amounts of hardware. The hardware complexity is proportional to the product of the number of PAA elements (K) and the number of PAA steering angles (R). Since K and R are in the [10.sup.2] to [10.sup.4] and [10.sup.2] to [10.sup.3] ranges, respectively, hardware compressive TTD architectures must be used that reduce the overall hardware with respect to both K and R. Such architectures(11,16,17) have been developed that employ binary FO delay lines (BIFODEL) in conjunction with optical WDM.
Hardware-Compressive WDM TTD
Hardware compression with respect to the number of PAA steering angles R is most efficiently accomplished using BIFODELs.(9) In a BIFODEL, the optical signal is optionally routed through N fiber segments whose lengths increase successively by a power of two. The various segments are addressed using a set of M 2 x 2 optical switches. Since each switch allows the signal either to connect or bypass a fiber segment, a delay T that can take any value may be inserted in increments of [Delta]T, up to the maximum value of ([2.sup.M] - 1) [Delta]T. Therefore, the complexity of the BIFODEL grows slowly with increasing R, in fact as [log.sub.2]R, and so only a small amount of hardware is needed even for large R.
The BIFODEL concept alone does not solve the overall hardware complexity problem because a K-element PAA requires K different BIFODELs. Further hardware compression can be accomplished(11,16,17) with respect to the number of PAA elements K by employing antenna partitioning in conjunction with WDM. Assuming that a K-element linear one-dimensional PAA is partitioned into E sets of N elements each, such that K = N x E, the delay required by the i-th element of the j-th set is equal to the delay of the i-th element of the first or reference set (RS) plus a bias delay. This bias delay depends only on j and not on i, and thus is common to all the elements of a given set. Therefore, if partitioning is used, the total number of different types of BIFODELs is reduced to N + E, that is, N for the RS plus E for the bias delays, as compared to N x E for the non-compressed implementation. In this case, the overall hardware complexity [M.sub.c] (with N = E = [square root of] K) is
[M.sub.c] =([log.sub.2]R) x (2 [square root of] K - 2) (1)
compared with [M.sub.c] = R x K for the straightforward implementation.
The partitioned PAA concept can be uniquely implemented optically using WDM, and is reversible, that is, the same hardware is used for the transmit and receive modes. A block diagram of a 16-element transmit WDM TTD architecture is shown in Figure 1. In the transmit mode, the input radar signal modulates the intensity of N laser diodes (LD) each at a different optical wavelength [[Lambda].sub.i]. The optical signal at wavelength [[Lambda].sub.1] propagates through a nondelayed path, whereas the remaining (N-1) optical signals at wavelengths [[Lambda].sub.i], i = 2,3, ... N, propagate through (N-1) different BIFODELs. The optical signals exiting the BIFODELs and the nondelayed path are multiplexed (MUX) and subsequently divided into E channels via an E-channel optical divider. (N-1) of the driver outputs drive (N-1) different bias BIFODELs, the outputs of which are subsequently demultiplexed (DMUX). The Nth output of the divider, that is, the one with no bias delay, is also demultiplexed, and the N resulting optical signals contain the N progressively delayed signals required for the RS that requires no bias delay. The outputs of each remaining DMUX contain a similar set of signals, but which are further delayed via the bias BIFODELs, and correspond to a different PAA set. Similar wavelength outputs drive similar location elements in each set.
The identical system is used for the receive mode, but in reverse, with the output of each PAA element driving a LD of a different wavelength. Elements with similar locations in different sets drive LDs of the same wavelength. For each PAA set, the LD outputs are multiplexed and drive a bias BIFODEL. The outputs of the bias BIFODELs are combined via the E-channel optical combiner, the output of which is subsequently demultiplexed, with each channel driving an RS BIFODEL. The outputs of the RS BIFODELs are fed to a combiner, the output of which provides the desired vector sum. Although the WDM TTD architecture is described with reference to a one-dimensional linear PAA, the identical architecture can be used for two-dimensional PAAs.(16) In all cases, the number of required BIFODELs increases slowly with increasing K as 2 ([square root of] K-1) (assuming N = E = [square root of] K), and thus only a small number of BIFODELs are needed even for large K.
Further hardware optimization of the TTD system is possible depending on the actual PAA application.(17) For example, consider one-dimensional PAAs used for surveillance radars operating in the UHF-, L- and S-bands. The physical length of most PAAs used in these applications results in maximum RS BIFODEL bit delays of the order of 2 ns. For such delays good quality coaxial cables and/or microstrips can be used without serious differential attenuation or delay (or phase) dispersion effects as a function of frequency. In addition, low cost 1 x 2 GaAs FET switches are available that operate over the S-band with low insertion loss ([is less than] 0.5 dB) and have a response that is flat (to better than [+ or -]0.05 dB) over the 0.2 to 3.5 GHz band. Therefore, for these applications, the RS BIFODELs can be implemented using all-electronic techniques, that is, electronic binary delay lines (DiBi).
In a DiBi, each time delay bit can be implemented via two back-to-back 1 x 2 switches connected via two different delay microstrips. The advantages of DiBis over BIFODELs for implementing the RS portion of the system include lower cost ([approximately] 2 orders of magnitude less per delay line, mostly due to reduced switch cost), the potential for RS delay implementation in integrated form using GaAs MMIC and/or wafer-scale techniques and smaller size. Use of DiBis for the RS delay and BIFODELs for the bias delays results in a hybrid TTD WDM architecture. In this architecture, the radar signal is applied to N-1 different DiBis as well as a nondelayed path. The N output electronic signals are then used to modulate the intensity of N LDs each at a different optical wavelength [[Lambda].sub.i]. The N optical signals are then multiplexed into one optical channel, which is then divided into E channels and each is applied to a BIFODEL. The hybrid architecture has several practical advantages. It uses fiber optics only where standard low cost micro-wave electronic techniques cannot perform, and it preserves the unique features of optics, that is, it uses WDM for implementing the PAA partitioning, and optical fiber for implementing long delays. It is not necessary to implement all of the bits of the RS delay lines in the electronic domain. As many bits as possible can be implemented electronically and then reverted to fiber-optic delays prior to WDM. This approach allows the hybrid scheme to be used for large PAAs for which the sole use of DiBis in the RS level may not be possible. Since both the DiBis and BIFODELs are reversible, the hybrid architecture is also reversible.
Prototype System Development
To evaluate the performance of the hybrid WDM TTD system, a prototype 16-element transmit TTD system was developed and demonstrated with a BW of 0.35 to 2.1 GHz, a scan angle of [+ or -]45 [degrees] with delay resolution of 6 bits (1.31 [degrees]), reconfiguration time of the order of ms, RMS phase error of [is less than] 5 [degrees] over the full band, and element-to-element amplitude uniformity of better than [+ or -]0.75 dB. The [+ or -]45 [degrees] scan angle specification was dictated by the characteristics of the radiating elements of the available PAA rather than by the TTD system. The actual used PAA restricted the BW to 0.7 to 1.4 GHz and had an element spacing of 10.707 cm. For this PAA, the delay ranges, as determined by the LSB and MSB delay bits of the delay lines, were 8.2 to 260.7 ps, 16.3 to 521.4 ps, and 24.4 to 782.1 ps for the three DiBis, and 32.6 to 1042.7 ps, 65.2 to 2085.5 ps, and 97.8 to 3128.2 ps for the three BIFODELs. To implement negative scan angles, the antenna was time-biased so it is steered at -45 [degrees] when the TTD network provides zero delay. When the TTD provides the maximum delay, the PAA is steered at +45 [degrees]. This performance is accomplished by inserting a linearly-varying but fixed time-bias into the feeds to the various elements.
Electronic Delay Line Development
Three different generations of electronic delay lines (DiBi) have been developed using a combination of proprietary and well established microwave fabrication techniques. They are of low cost ([is less than] $250 for a six-bit DiBi), bidirectional, and with the latest generation, have 0.5 dB BW covering the 0.7 to 1.4 GHz range and a [+ or -]0.25 dB peak-to-peak (better than 0.1 dB RMS) matched response for the various delay combinations. Figure 2 shows a latest generation prototype six-bit DiBi with an LSB = 24.4 ps and an MSB = 782.1 ps at 0.7 to 1.4 GHz. These DiBis are high speed delay lines, and their program can be changed every 15 ns. The worst case delay deviation from the design values is [is less than] [+ or -]4 ps peak-to-peak (or [is less than] [+ or -]2 [degrees] of phase at 1400 MHz) and is sufficient for many high performance radar applications. The delay error has a zero mean uniform distribution with a standard deviation of 1.45 ps (or 0.72 [degrees] of phase at 1400 MHz), and is independent of the delay range. The measured SNR is 177 dB/Hz and the measured spurious-free dynamic range is 135 dB/[Hz.sup.2/3]. The insertion loss is -6 dB and the noise figure is 6 dB. No crosstalk could be measured using a time-domain technique with a sensitivity of -65 dB. The theoretical cross-talk, estimated to be -90 dB, is due mostly to the switch crosstalk.
Fiber-optic Delay Line Development
Two generations of fiber-optic delay lines (BIFODELs) have been developed. The BIFODELs use customized piezomechanical 2 x 2 optical switches that have no differential delay between the two switch paths. Their optical insertion loss is [is less than] 1 dB ([approximately]0.7 dB average), their optical crosstalk is [is less than] -65 dB (or -130 dB RF), and their rise time is 1 ms, which allows switching speeds as low as 3 to 4 ms. Such switches are satisfactory for demonstration purposes, and furthermore, they are sufficiently fast for some UHF- and L-band PAA applications. However, for most high performance radar applications, the switching speed must be reduced by at least an order of magnitude, which is a difficult task for piezomechanical switches. For this purpose, other switch technologies, such as ferroelectric liquid crystal switches(18) with rise times of 150 [[micro]seconds] or integrated optical switches, must be employed.
The BIFODELs are packaged on a 10 x 14 [in.sup.2] plexiglass board that also houses the DMUX. The overall dimensions of the BIFODEL are determined by the number and dimensions of the switches, and by the minimum bend radius of the fiber pigtails necessary to avoid loss of light between switches. The BIFODELs are bidirectional and their delays are set differentially. Overall, the time delay performance is similar to that of the DiBis. The worst case measured delay deviation from the design value is [is less than] [+ or -]2.5 ps peak-to-peak or [is less than] [+ or -]1.25 [degrees] of phase at 1400 MHz. The delay error is similar to that of the DiBis and has a uniform distribution with a zero mean and a standard deviation of 1.42 ps or 0.71 [degrees] of phase at 1400 MHz. The crosstalk is lower than the measurement limit of -110 dB, which is to be expected since the used switches have crosstalk levels better than -130 dB.
Overall, the time delay performance of the BIFODELs is found to be similar to that of the DiBis, which is an interesting result in view of the fact that completely different technologies are involved. Both devices are bidirectional but the BIFODEL requires much larger volume. However, the BIFODEL can implement much longer delays (for example, hundreds of [[micro]seconds]) and is transparent to microwaves, whereas the DiBi is limited to a few ns and a few GHz of bandwidth. However, the most striking difference is their cost, which is under $250 for a DiBi vs. over $7500 for a BIFODEL, the major factor being the cost of the switches.
Fiber-Optic Link and Fabrication
To provide the four different wavelengths, commercially available distributed feedback (DFB) LDs have been used with wavelengths of 1301, 1308.7, 1314.3 and 1321.6 nm. Each LD is packaged in a module that also contains a power regulator, a soft-start circuit, a bias supply regulated by a closed-loop amplifier fed by a signal derived from the integral pin diode of the LD package, a thermistor and a closed loop to control the integral LD thermoelectric cooler, and a microstrip matching network for the LD designed to flatten the response of the LD-DET combination. The amplitude response of a typical LD module-output detector (DET) combination is flat to within [+ or -]0.2 dB over the 0.7 to 1.4 GHz band, whereas the delay dispersion is flat to better than [+ or -]3 ps over the 0.7 to 1.4 GHz band. These results demonstrate that optimized FO links can achieve performance comparable to that of the DiBis. When the same DET is used, the responses of the four FO links are matched to within [+ or -]0.25 dB over the full 0.7 to 1.4 GHz band. However, in practice, this matched performance deteriorates somewhat because different DETs are used for the 16 output ports, and because a transmit module with 60 dB of gain has been incorporated at the output of each detector. The frequency response of the DET-transmit output module is flat to 0.5 dB over the 0.7 to 1.4 GHz range of interest. This performance is compatible with that of the DiBis and the FO links, and does not degrade the overall system frequency response. The phase is flat to within [+ or -]5 [degrees], which is worse than the phase response of the DiBis by a factor of [approximately]2.5, and is the limiting factor of the system phase response.
WDM TTD Prototype System Integration
The prototype TTD system was assembled in three packages, the delay generation system, the FO feed together with the DET-gain block module for each element, and the computer control and power supply. Figure 3 shows a photograph of the delay generation system, the FO feed and the 16 output modules of the packaged TTD system. The input signal to the TTD system is fed to a high performance 0.35 to 2.1 GHz 1:4 RF divider, the outputs from which drive the three DiBis and the zero path module. The outputs from the DiBis are connected to the LD modules via delay-equalized cables. To sum and broadcast the outputs from the four LDs to the BIFODELs, a 4 x 4 wavelength independent coupler was used that performs the combined functions of a four-channel MUX and 1-to-4 divider. The optical signals from the BIFODELs are demultiplexed via four bidirectional DMUXs, which have a 90 percent optical bandwidth of 0.48 nm and an insertion loss of [is similar to]3 dB. Their cross-talk level is less than -75 dBc, most of which is due to RF pick-up rather than optical leakage. The integrated system is computer controlled using menu-driven software.
Laboratory Testing and Evaluation
The integrated TTD system has been subjected to several tests in order to evaluate the system performance prior to the antenna range experiments. The overall delay generation has been verified by accurately measuring the relative output delays for each of the 16 output elements and for each steering angle. Figure 4 shows an example of the design and the measured delays for the 16 output elements as a function of the six-bit (63 combinations) switch position. The switch position relates to the steering angle with a 1.31 [degrees] resolution. These data are virtually identical for any frequency in the 0.7 to 1.4 GHz band, and are also similar to those obtained by propagating pulses of 500 ps half-width. The standard deviation of the worst case time error over the 16 output channels is 9.8 ps over the full band, whereas the best case standard deviation is 4.03 ps, that is, 4.8 [degrees] and 2 [degrees] at 1.4 GHz. These standard deviations clearly demonstrate that sufficient system-level time delay performance is possible without error correction.
The 3 dB frequency response of the full system, that is, any of the 16 output ports, covers the 0.35 to 2.1 GHz band, with the lower frequency being limited by the input 1:4 RF divider (for the DiBis) and the upper frequency being limited by the DET and the output amplifiers. Over the 0.6 to 1.5 GHz band, the response of any of the 16 outputs is flat to [+ or -] 0.5 dB. The SNR available at the output of each of the 16 ports is determined primarily by the maximum SNR possible from the LDs, which are 144.3, 150.1, 150.2 and 152.2 dB/Hz (the limiting factors being the LD relative intensity noise levels), and the optical insertion loss. Given that the measured average optical loss is 14 dB, the average system SNR is 120 dB/Hz. The SNR improves to 140 dB/Hz if a transimpedance amplifier (TRAM) is used at each output module. However, matching the responses of the output TRAMs is a significantly more difficult task than matching conventional voltage amplifiers. The described performance applies to a transmit system where the power of each LD is divided into [square root of K] channels. For the receive TTD system, the LD signal is not divided and the RF signals from all the LDs are coherently added so that significant improvements (at least 12 dB for the 16-element system) are possible.
Antenna Range System Demonstration
The PAA used for the TTD demonstration consists of 16 columns, as shown in Figure 5. Each column contains two radiating elements, which are vertically polarized, reduced depth notch radiators, 0.3[Lambda], deep ([is less than] 3"), and simultaneously broadband at 850 to 1400 MHz (1 dB BW) or 700 to 1500 MHz (3 dB BW). The antenna was mounted 12 feet above ground on a pedestal that consists of a three-axis positioner, which can turn the antenna [+ or -]180 [degrees] horizontally and vertically. The TTD system was placed inside a sealed container to protect it from rain. The receive modules were connected directly to the input SMA connectors of the 16 PAA elements. A reference notch element mounted centrally on top of the PAA provides phase and amplitude references for the diagnostic software. A 2 x 4 notch element PAA (similar to the 2 x 16 element PAA used with the TTD system) was used as a receive antenna and was located about 200 feet down range from the antenna.
For recording an antenna pattern, the TTD system switches were set to the desired steering angle; a CW signal in the 0.6 to 1.5 GHz band was generated in the control room and fed to the input of the TTD system; the transmit antenna was rotated horizontally by [+ or -] 80 [degrees]; and the signal from the receive antenna was recorded as a function of the transmit antenna rotation angle. For each antenna pattern, a computer tuning program (MTUNE) was used to measure the amplitude and phase distribution over the 16 elements. MTUNE has been used for antennas with ultra-low sidelobes down to -55 dB. These pattern data are reverse transformed back to the aperture and a distribution is computed. From this distribution, an error plot of the amplitude/phase compared to the theoretical distribution is generated. The use of MTUNE proves extremely valuable because it helped identify various problems at the antenna-receiver module interface, for example, mismatches due to humidity, shorted output amplifiers and mutual coupling between elements.
The prototype WDM TTD system was tested successfully at the antenna range over a period of two weeks under various temperature and humidity conditions. No effects attributable to humidity variations were observed. However, [+ or -]1.3 dB gain variations were observed as the temperature varied, probably due to changes in the response of the DMUX devices with temperature. The main test involved the demonstration of the antenna beam stability as a function of frequency, which is the key characteristic of the TTD operation. For this purpose, the antenna was steered at a specific angle, the RF input frequency to the TTD system was changed in discrete steps over the antenna BW, and the resulting antenna pattern was recorded for each angle-frequency combination over a [+ or -]80 [degrees] scan. Figure 6 shows of some of the recorded squint-free antenna patterns obtained. Twelve superimposed antenna patterns are shown for -43 [degrees], 0 [degrees] and +45 [degrees] and for frequencies of 600, 900, 1200 and 1500 MHz. These specific angle and frequency settings were chosen because they cover the full scan range of the TTD prototype system and the maximum possible BW of the antenna. In all cases, stable, squint-free antenna patterns were recorded regardless of the operating frequency, which unambiguously demonstrates the TTD nature of the prototype system. Such patterns are not possible with phase steering. Similar results were obtained over the full range of steering angle covered by the TTD system and for any frequency covered by the BW of the antenna.
In another experiment, all even numbered PAA elements were switched off in order to double the element spacing and reduce the mutual coupling between adjacent PAA elements. This mutual coupling was caused by the unavailability of broadband RF isolators at the time of the antenna range experiments. Using this eight-element PAA arrangement, various antenna patterns were recorded with input frequencies restricted to the 600 to 700 MHz frequency band. Figure 7 shows theoretical and experimental patterns for an input frequency of 700 MHz. The measured sidelobes are within 1.8 dB of the theoretical sidelobes, which proves that no serious errors were produced by the TTD system. Similar behavior was observed for various steering angles and frequencies in the 600 to 700 MHz range. Furthermore, during these measurements, deep, that is, [is less than] -40 dB, nulls in the antenna patterns were frequently observed. In view of the small antenna size used in these experiments, such deep nulls are associated with small time errors, which is consistent with the results obtained in the laboratory time error tests.
Experiments were also performed to verify the angular resolution of the TTD system. For these experiments, the even-numbered elements of the 16-element antenna with a fixed input RF frequency of 700 MHz were used and a total of 63 patterns, one for each binary combination of the six-bit switches, were recorded. Examples of such patterns demonstrate the quantized (in binary form) resolution of the antenna, as shown in Figure 8. The steering angle sequentially doubles as the TTD program is switched in turn from 000010 to 000100 to 001000 to 010000. These experiments verify the expected system resolution (1.31 [degrees]) and the scan angle range ([+ or -]45 [degrees]) of the system.
The development, testing and antenna range demonstration of a hybrid 16-element transmit 0.35 to 2.1 GHz six-bit WDM TTD system has been described. The system is based on a hardware-compressive architecture. The integrated WDM TTD system was successfully demonstrated at an antenna range using a 2 x 16 element broadband phased-array antenna. Stable, squint-free antenna patterns have been obtained over the full 0.6 to 1.5 GHz antenna bandwidth and over the full TTD system scan range of [+ or -]45 [degrees]. Overall, the measured system performance is close to that predicted by the design, and satisfies the requirements for various scenarios of phased-array antennas used in radar and communications applications.
This work was performed under Westinghouse IR&D and under contract F30602-91 -C-0111 from Rome Laboratory. The authors thank B. Hendrickson and J. Hunter of Rome Laboratory for their support and encouragement of this work.
1. L. Cardone, "Ultra-Wideband Microwave Beamforming Technique," Microwave Journal, April 1985, pp. 121-131.
2. D. Curtis, "Frequency Domain Analysis and Performance of a True-Time Delay Fiber Optic Beamforming Network for Array Antennas," Proceedings SPIE, Vol. 1217, 1990.
3. W. Ng, et al., "The First Demonstration of an Optically-Steered Microwave Phased Array Antenna using True-Time Delay," Journal of Lightwave Technology, Vol. 9, No. 9, 1991.
4. G.A. Koepf, "Optical Processor for Phased Array Antenna Beam Formation," Proc. SPIE, Vol. 477, 1984.
5. P.R. Herczfeld and A.S. Daryoush, "Fiber-Optic Feed Network for Large Aperture Phased Array Antennas," Microwave Journal, August, 1987.
6. E.D. Toughlian and H. Zmuda, "A Photonic Variable RF Delay Line for Phased Array Antennas," J. Light Technol., Vol. 8, No. 12, 1990.
7. R.D. Esman, et al., "Microwave True-Time Delay Modulator using Fiber-optic Dispersion," Electronic Letters, Vol. 28, No. 20, 1992.
8. R. Soref, "Optical Dispersion Technique for Time-Delay Beam Steering," Appl. Opt., Vol. 31, No. 35, 1992.
9. A.M. Levine, "Fiber-Optic Phased Array Antenna System for RF Transmission," US Patent No. 4028702, 1977.
10. R.A. Soref, "Programmable Time-Delay Devices," Applied Optics, Vol. 23, No. 21, 1984.
11. A.P. Goutzoulis and D.K. Davies, "Hardware Compressive Two-Dimensional Fiber-optic Delay Line Architecture for Time Steering of Phased Array Antennas," Appl. Opt. 29, 1990.
12. D. Dolfi, et al., "Two-Dimensional Optical Architecture for True-Time Delay Beam Forming in a Phased Array Antenna," Opt. Lett., Vol. 18, No. 4, 1991.
13. C.T. Sullivan, et al., "Switched Time Delay Elements Based on AIGaAs/GaAs Optical Waveguide Technology at 1.32 [[micro]meter] for Optically-Controlled Phased Array Antennas," Proceedings of SPIE, Vol. 1703, 1992.
14. G.A. Magel and J.L. Leonard, "Phosphosilicate Glass Waveguides for Phased-Array Radar Time Delay, Proc. SPIE, Vol. 1703, 1992.
15. W. Ng, et al., "GaAs Optical Time-Shift Network for Steering a Dual-Band Microwave Phased Array Antenna," Proc. SPIE, Vol. 1703, 1992.
16. A.P. Goutzoulis and D.K. Davies, "All-Optical Hardware-Compressive Wavelength Multiplexed Fiber-optic Architecture for True-Time Delay Steering of Two-Dimensional Phased Array Antennas," Proc. SPIE, Vol. 1703, 1992.
17. A.P. Goutzoulis, et al., "Hybrid Electronic Fiber-optic Wavelength Multiplexed System for True Time-Delay Steering of Phased Array Antennas," Opt. Eng., Vol. 31, No. 11, November 1992.
18. M.R. Meadows, et al., "Electro-Optical Fiber-Optic Switches having Low Cross-talk," Conference of Optical Society of America, San Jose, CA, 1991.
Akis P. Goutzoulis received his BS degree in physics from the University of Ioannina, Greece, in 1978, his MS degree in optoelectronics from the University of Essex, England, in 1979, and his PhD degree in electrical engineering from Carnegie-Mellon University in 1983. From 1976 to 1977, he performed research in digital electronics in the Digital Laboratory of Democritos NRC in Athens, Greece. Since 1983, he has been with the Westinghouse STC, where he has been involved in the research of optics, acousto-optic devices, acousto-optic and electro-optic signal processing, optical computing, optical communications, wideband fiber-optic distribution for control and true-time delay steering of phased array antennas, and microwave and mm-wave analog Fiber-optic links. Goutzoulis holds 12 patents. He is a member of OSA, SPIE and IEEE.
D. Kenneth Davies received his BSc and PhD degrees in physics from the University of Wales (UK) in 1957 and 1960, respectively. Following a two-year fellowship at the University College of Swansea (UK), he joined the Westinghouse R&D Center in 1962, where he has been engaged in research in the areas of atomic physics, spectroscopy, vacuum discharges, gaseous electronics, optical computing and optical signal processing. His current interests include the development of miniature ion mobility gas sensors, high-fidelity microwave fiber-optic links and fiber-optic true-time delay systems for phased array antennas. Davies holds four patents and is a fellow of the Institute of Physics, a member of APS.
John M. Zomp is an engineering specialist at the Westinghouse R&D Center. He joined Westinghouse in 1966 after having worked at the University of Pittsburgh's Physics Department on instrumentation for the cyclotron and Van de Graft accelerators. His many activities at the R&D Center include circuit design, directional planning, instrumentation consulting, and managing an electronic model shop. Zomp has numerous patents and received a Westinghouse Engineering Achievement Award in 1983.
Peter D. Hrycak received his BSEE degree, magna cum laude from Stevens Institute of Technology in 1973, and his MSEE from the University of Maryland in 1980. He joined Westinghouse Electric Corporation in 1973. He has been involved in design of advanced antenna systems. In 1987, he was promoted to the position of supervisory engineer of the Advanced Aperture Systems group. His primary interest is in pushing the state-of-the-art of array systems. Hrycak is a senior member of the IEEE, the APS, and the Systems Society, as well as past chairman of the Baltimore section.
Andrew H. Johnson received his BS degree in 1961 in electrical engineering from the University of Illinois. He joined Westinghouse Electric Corporation in 1961. His primary fields of interest are radiator design and advanced antenna pattern measurement techniques. Johnson is a member of Eta Kappa Nu and is a registered professional engineer in Maryland.
|Printer friendly Cite/link Email Feedback|
|Author:||Goutzoulis, Akis; Davies, Ken; Zomp, John; Hrycak, Peter; Johnson, Andy|
|Date:||Sep 1, 1994|
|Previous Article:||Determination of E(sub.eff) for printed structures on electrically thin substrates.|
|Next Article:||The ambidextrous antenna.|