The basic operation principles of semiconductor laser diodes are described and key characteristics and current status of commercial laser diodes that are relevant to optical signal processing applications are discussed.

Future electromagnetic environments will have much higher signal densities than current system processors are capable of detecting, sorting, and classifying. New channelizer system processors utilizing acousto-optical Bragg cells followed by advanced detector arrays and microelectronic signal preprocessing circuits to minimize the input signal rate to the digital processor have the potential to meet future system processing needs. Detector considerations for the required, new channelizer system processors are discussed.

Detectors for acousto-optic signal processing applications require fast transient response as well as wide dynamic range. There are two major choices of detectors: conductive or integration mode. Conductive mode detectors have an initial transient period before they reach then' i equilibrium state. The duration of 1 his period is dependent on light level as well as detector capacitance. At low light levels a conductive mode detector is very slow; response time is typically on the order of milliseconds. Generally. to obtain fast transient response an integrating mode detector is preferred. With integrating mode detectors. the dynamic range is determined by the charge storage capability of the tran-sport shift registers and the noise level of the image sensor. The conventional net hod used to improve dynamic range is to increase the shift register charge storage capability. To achieve a dynamic range of fifty thousand assuming two hundred noise equivalent electrons, a charge storage capability of ten million electrons would be required. In order to accommodate this amount of charge. unrealistic shift registers widths would be required. Therefore, with an integrating mode detector it is difficult to achieve a dynamic range of over four orders of magnitude of input light intensity. Another alternative is to solve the problem at the photodetector aml not the shift, register. DALSA's wide dynamic range detector utilizes an optimized, ion implant doped, profiled MOSFET photodetector specifically designed for wide dynamic range. When this new detector operates at high speed and at low light levels the photons are collected and stored in an integrating fashion. However. at bright light levels where transient periods are short, the detector switches into a conductive mode. The light intensity is logarithmically compressed into small charge packets, easily carried by the CCD shift register. As a result of the logarithmic conversion, dynamic ranges of over six orders of magnitide are obtained. To achieve the short integration times necessary in acousto-optic applications. t he wide dynamic range detector has been implemented into a tapped array architecture with eight outputs and 256 photoelements. Operation of each 01)1,1)111 at 16 MHz yields detector integration times of 2 micro-seconds. Buried channel two phase CCD shift register technology is utilized to minimize image sensor noise improve video output rates and increase ease of operation.

Acousto-optical signal processors require diffraction limited images positioned to accuracies of one percent or less. Wavefront tolerances are twentieth wave. On the other hand, the light is usually monochromatic. Fields of view are modest and, in fact, negligible where fast (N.A.-55) optics are required, e.g., laser diode collimators, objective lens are, typically, scan lenses. Some aberrations are not critical but field curvature must be corrected and the system must be aplantic.

The performance of optical processors, particularly crosstalk and dynamic range, normally does not measure up to what is predicted by first order optical design, ray tracing, and diffraction calculations. Contributing factors such as detector array and light source characteristics, discussed elsewhere by other authors, combine with optical performance factors such as large-and small-angle scatter and multiple reflections to determine the actual system performance. A stray light computer code, ASAP/RABET, has been used to model several simple configurations found in optical processors. Scattering of samples of material was measured using a low angle scatterometer in transmission (BTDF), and measured longscan surface roughness profile data (WYKO TOPO). The results including the stray light effects are compared with normal calculations (CODEV) and also with some measured performance data.

The results of a computer model of the far field diffraction pattern of the Texas Instruments cantilever beam inverted cloverleaf deformable mirror device (DMD) are presented. Both amplitude and phase in the Fourier transform plane are studied using a variety of input DMD phase objects. Changes in the Fourier plane phase and amplitude produced by changes at the input (including random leaflet-to-leaflet input phase errors) are investigated. The phase in the Fourier plane was found to be very sensitive to even slight variations in the deflections of the individual mirror leaflets. This could limit its usefulness as an input device.

Optically-addressed spatial light modulators (OASLMs) utilizing a hydrogenated amorphous silicon photodiode as the photosensor and a ferroelectric liquid crystal as the modulator show great promise in high-speed, high-contrast spatial light modulation. We present the initial design, modelling, and performance of such devices. The devices are driven by a square-wave voltage, such that write and read operations take place under reverse bias, and an erase operation occurs under forward bias. Under continuous write-beam illumination the optical modulation is determined by the photoconductivity of the photosensor under forward bias. During chopped write-beam illumination the modulation is produced by the photocurrent produced under reverse bias. We find an increase in the contrast with reduced bias voltage, which is explained in terms of a series capacitance effect. Several polarization-based logic gates are described which utilize a single OASLM devices.

The use of monocrystalline gallium arsenide as a photoaddressing medium for a nematic liquid crystal spatial light modulator is reported. A GaAs-addressed nematic SLM is demonstrated with a limiting resolution of 37 1p/mm over a 30-mm aperture and a 100-Hz frame rate. The influence of the GaAs thickness on the SLM resolution is analyzed.

Four spatial light modulators (SLM's) with varying efficiency light blocking layers are used to demonstrate the effect of optical isolation on the devices' performance. It is shown that incorporation of high attenuation blocking layers can improve most aspects of the SLM's behaviour, yielding a 300 fold increase in gain over the the device without a blocking layer, and also improve the device resolution to give a 50% MTF at 50 line pairs/mm. A non-ambiguous method of determining the devices sensitivity is also presented along with a simple formula relating the various intensity requirements for a particular application to the gain and light blocking requirements of the SLM.

A detailed characterization of crosstalk was carried out for two 32-channel. Te0₂ Bragg cells: one operated at 95 MHz, the other at 400 MHz. A survey showed that the two cells were state-of-the-art devices. Crosstalk was visualized and measured radiometrically in a Schlieren imaging setup. Both qualitative and quantitative results were obtained. Results showed that, when neighboring channels were activated, acoustic crosstalk due to beam spreading was the major source, especially when all the channels are driven by coherent rf signals. Acoustic crosstalk varies from zero at the transducer to about 10 dB below the maximum beam intensity. Radio frequency crosstalk was also measured. It amounted to -25 dB in the 95-MHz cell and -35 dB in the 400-MHz cell. Possible methods of further reducing the crosstalk were given and suggestions for future work were made.

Applications of GHz bandwidth acousto-optic devices operating at 630nm and 830nm seem to be ideal for optical signal processing, scanners, deflectors and frequency shifters. One of the major applications are for RF spectrometers in early warning radar receivers and auto-correlators.

Bragg cells using the slow-shear wave propagating along the [110] direction in Te0₂ have become widely utilized in the optical signal processing field. Devices with aperture lengths up to 100 gsec have been fabricated using this mode. The center frequencies of such devices have ranged from 50 to 90 MHz with bandwidths ranging from 20 to 60 MHz. For many signal processing applications, however, significant improvement in the Bragg cell performance can be obtained if either an optically rotated or an acoustically rotated variation on the standard crystal cut is used instead of the typical crystal orientation. These slightly varied orientations can result in significant enhancement in diffraction efficiency and dynamic range as well as in allowing the Lre of linear polarization on the input beam. The work presented examines in detail the various physical properties of these special crystal cuts. The paper will also present guidelines as to when one of these crystallographic rotations will yield significant advantage over the standard orientation.

Design considerations for a compact acousto-optic triple-product-processor are presented. This system, using Te0₂ Bragg cells operating in the slow shear wave mode, does not require the use of anarrnorphic (cylindrical-spherical lens) optics. The designs for the two Bragg cells are carried out in detail. Measurements of their performance are also reported. Good agreement was obtained between theory and experiment. The bandwidth of the overall system is 47 MHz, each Bragg cell having a time-bandwidth product of 470.

The design of optical activity-based long delay Hg₂Cl₂ Bragg cells is discussed, and the latest experimental data for crystal growth, crystal quality, transducer response and Bragg cell performance are presented.

A general method for calculating the scattering amplitude of multifrequency acousto-optic diffraction is established. The method is based on counting allowable Feynman diagrams. It is found that the ratio of the number of Feynman diagrams allowable in the Bragg regime (isotropic, birefringent, and degenerate) to that in the Raman-Nath regime is independent of the total number of different acoustic frequencies, being a function only of the order of the Feynman diagram and the diffraction order of the final state. A general expression for this ratio is obtained. With this as a basis, complete perturbation solutions of the scattering amplitude can be obtained for any final state, any number of acoustic frequencies, and any kind of multifrequency acousto-optic diffraction. The theory is verified by comparing with theoretical results obtained previously and with experiment results.

The results of our investigation of the applicability of optical processing to Adaptive Phased Array Radar (APAR) data processing will be summarized. Subjects that are covered include: (i) new iterative Fourier Transform based technique to determine the array antenna weight vector such that the resulting antenna pattern has nulls at desired locations, (ii) obtaining the solution of the optimal Wiener weight vector by both iterative and direct methods on two laboratory Optical Linear Algebra Processing (OLAP) systems, and (iii) an investigation of the effects of errors present in OLAP systems on the solution vectors.

A novel broadband beamformer based on space-integrating, interferometric multichannel acousto-optic correlators was proposed. Both analysis and experiments were carried out to validate the concept. Simultaneous beamforming and pulse compression were demonstra'ted by an experimental simulation with a 126-bit PN coded signal of 50-MHz bandwidth. Analysis also showed that the system was suitable for multiple beamforming. The responses of the system to narrow (cw) and wideband (random noise) jammers, both on-boresight and off-bore-sight, were investigated experimentally. Results showed that, against these interferences, the system was quite robust due to the processing gain, even for the worst case where the jammers was on-boresight and its frequency coincided with the signal carrier. However, the system's sensitivity was reduced by a factor of four for the wideband jammer and a factor of ten for the narrow band jammer. The dynamic range of the system was measured to exceed 20 dB, only to be limited by detector saturation and Bragg cell non-linearity. The system could be made adaptive.

This paper describes realizable acousto-optic configurations for the implementation of phased array beamformers. The simple system is capable of generating a distribution of wideband radio frequency (RF) signals with controlled phases. A laboratory demonstration model of the acousto-optic beamformer was built using a pair of Bragg cells and a 10-channel fiber-detector array. Experimental results show that a large phase excursion is obtainable simply by tuning the frequency of the control signal.

Signal excision will be defined, as well as some of its practical advantages. An acousto-optic signal excisor based on an interfero-metric spectrum analyzer will be presented. This system is capable of notch filtering 1 MHz channels across a 200 MHz RF bandwidth. Experimental results will be discussed,as well as comments on future system goals and required technology developments.

This paper discusses system aspects of adaptive filtering applied to narrowband interference rejection for direct sequence spread spectrum systems. Removal of strong CW interference can greatly improve the operation of downstream subsystems configured for detection of wideband signals. At the same time, the desired signal may undergo distortion due to the filter. The mechanism for that distortion is introduced, and the effects on system operation are shown.

This paper describes laboratory tests performed on a single channel multiplicative space integrating acousto-optic correlator/receiver. These tests were performed to demonstrate the feasibility of using a similar multi-channel system to perform synchronization and demodulation of two types of spread spectrum signals. Laboratory results demonstrating coherent and noncoherent correlation postprocessing techniques are presented for both types of spread spectrum signals.

A multichannel acousto-optic signal correlator based on optical joint transform correlation (OJTC) is presented. Parallel on-plane acousto-optic (AO) cells and a square law converter are used to produce cross-correlation of temporal-spatial signals. The AO cells are driven by frequency multiplexed temporal signals to produce spatially modulated signals. The joint Fourier transform of these spatially modulated signals is taken at the Fourier plane, where a square law detector converts the signals to a joint power spectra distribution. By simple coherent read-out, crosscorrelation of the input signals can be obtained at the output plane of the OJTC. Note that the proposed architecture can be packaged in a compact form. Since the multi-cell architecture can process broad band communication signals, we believe that the proposed technique can be applied to multi-frequency shift keying, as well as frequency hopped spread spectrum signals.

This paper is an experimental investigation of the performance of a real-time spread spectrum optical signal processor. The technique developed here is used to detect large time bandwidth product spread spectrum signals transmitted by the frequency hopping method and can yield self bit synchronization. This real-time optical system can generate the two dimensional correlation distribution between the reference data and the input data at the output plane of the processor. The conversion of the spread spectrum data from the time domain to the space domain is performed by acousto-optic Bragg cells and charged coupled devices. The two dimensional correlation between the input data and the reference data is obtained by joint transform correlation. The spread spectrum processor consists of a real-time optical correlator. Experimental results obtained by the real-time spread spectrum optical processor for various frequency hopped signals are presented. Various methods of converting the time data to two dimensional code patterns are described.

This paper describes an acousto-optic interferometer that is capable of measuring the instantaneous phases of wideband radio frequency (RF) signals. A novel homodyne detection concept was proposed in the implementation of the acousto-optic phase interferometer (AOPI). The AOPI was utilized to construct a channelized instantaneous frequency measurement (IFM) receiver. A breadboard of the AOIFM was built and tested. Experimental results of the breadboard are reported in this paper.

We present a theory of optimum coherence recovery applicable in computation-limited environments. We describe approaches for implementing coherence recovery employing two-dimensional Fourier transform acousto-optic architectures which afford very high throughput signal searches. The optimum one parameter, second order approximation to a small portion of a sinusoidally chirped sinusoid is a quadratic time transformation. The algorithm exploits the coincidence that the product of two chirps of nearly equal acceleration generates a Fourier kernel with an additional temporal phase which can exactly compensate a quadratically shifted sinusoid. Since for a large variety of functions the quadratic approximation is near optimum, such devices may have wide applicability for coherence recovery. A family of quadratic transformations may be applied, using a high throughput acousto-optic device, to effect recovery of an unknown periodicity in near real time. Alternatively, the technique may be employed to determine precisely the difference in time derivatives of frequency between two chirped signals. We describe our technique for implementing the coherence recovery algorithm in a prototype one-dimensional time-integrating Fourier processor. We have demonstrated analytically that the algorithm is realizable in a high bandwidth, two-dimensional hybrid (space-and time-integrating) system. High performance analog devices which effect coherence recovery have immediate relevance for detection of weak signals from electronic sources with phase modulation or sources undergoing acceleration, e.g., for gravitational wave experiments, and radio and X-ray searches for binary millisecond pulsars.

We describe heterodyne spectrum analyzers that operate with a significantly reduced number of photodetector elements. The technique is to decimate an N-element photodetector array by retaining only every M^th element. We time-share the remaining elements by scanning the spectrum across the decimated array. Each photodetector therefore produces, as a time sequence of samples, the spectral content of the received signal over its associated spatial frequency range.

Acousto-optic interferometric spectrum analysis performs coherent channelization and downconversion of a large number of channels within a wide frequency band. Each IF output requires a photodetector which is typically followed by a pre-amplifier, bandpass filter, and video logarithmic amplifier. Systems with 1000-2000 channels are exceedingly complex. This paper presents experimental results of an architecture which employs an additional Bragg cell to scan the spatial frequency plane across a decimated photodetector array, reducing the number of post detection circuits by effectively timesharing detector channels. The architecture is tested both as an interferometric-and cross-spectrum analyzer which facilitates angle of arrival measurement by preserving the phase difference between the two input signals.

We describe an acousto-optic signal processor which performs both pulse compression processing and Doppler filtering of pulse-Doppler radar signals. The processor combines a wideband space integrating acousto-optic pulse compression technique with a narrowband acousto-optic time integrating spectrum analysis technique to resolve target range and velocity simultaneously. Using a two-dimensional architecture, both pulse compression and Doppler filtering are performed in a potentially small, low power, package.

This paper describes a programmable acousto-optic processor that uses bipolar joint transform correlation for signal detection. The performance of the bipolar joint transform-based acousto-optic corre-lator is compared to that of the classical joint transform-based acousto-optic correlator in the areas of light efficiency, correlation peak to sidelobe ratio, auto-correlation bandwidth, and cross-correlation sensitivity.

The Wigner Distribution Function and Ambiguity Function are both two-dimensional time-frequency representations of a one dimensional signal. In spite of the similarity of the functional definition, however, the information is encoded in the two functions very differently. This paper reiterates the important traits of both and presents a laboratory setup using acoustooptic cells where both of the functions may be generated with no physical change to the hardware. The theoretical capabilities of such an optical system in terms of equivalent math operations is shown. Also a generalized version of the WDF which demonstrates the additional trait of being positive semi-definite is presented with the technique for adding this capability to the optical design.

Moire contours are seen on a curved surface viewed through a grating when the surface is illumnated with a matched grating. We have been able to replace both the projection and viewing physical gratings with gratings of variable spacing generated in acousto-optic cells, and have been able to observe and record Moire contours. The gratings are generated by amplitude modulating 70 MHz AO cells with a 1 to 10 MHz square wave. The gratings become visible on the curved object by strobing the illuminating laser beam with 50 nsec pulses by means of an AO modulator. Data is given showing the variation in grating spacing with AO modulation frequency. Figures are presented showing the variation of the Moire contour spacing with the AO modulation frequency and with target contour. The results for an angled flat plate target are compared to those in the literature.

"This material is based upon work supported by the National Science Foundation, under award number ISI-8660934. Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors; and do not necessarily reflect the views of the National Science Foundation."

An optical processor to solve partial differential equations for computational fluid dynamics applications is considered. This application is new and original for optical processors. The algorithms that are used are optical realizations of the Newton-Raphson method for nonlinear equations and a new optical LU direct decomposition and Gauss-Seidel iterative solution to the resultant linear algebraic equations. These algorithms are used to solve Burger's equation (a specific form of the momentum equation in fluid dynamics). The nonlinear equations provide 1-D velocity data at each time step. Simulation results of optical processing with these algorithms on computational fluid dynamics data is included.

In a Residue Number System, Second Factorization reduces modular multiplication, addition and subtraction to several miniature look up tables (LUT's). Optical implementation of Second Factorization for parallel matrix processing is investigated for large adaptive phased array radar applications. It is shown that a modified Sherman-Morrison method can be used with Second Facorization in a Quadratic Residue Number System, QRNS, to solve the adaptive beam forming problem. Second Factorization techniques are invoked to implement Quadratic RNS and General RNS arithmetic with small optical LUT's. Properties of the complex primes in QRNS/GRNS are discussed. It is shown that non-Archimedean primes as well as the complex factors of Archimedean primes can be used in a general residue number system (GRNS). Optical architectures for multiplication and addition (subtraction) are presented. Examples of these operations are illustrated. A photograph of a preliminary Mod (6+5i) matrix-vector processor is shown.

A novel system for solving systems of nonlinear equations is proposed. Two different algorithms are introduced. A speed analysis of the two different algorithms is presented and compared with the speed of their digital computer counter parts. A great advantage in speed is shown for large size problems.

Hardware and software design of the Bimodal Optical Computer (BOC) and its implementations are presented. Experimental results of the BOC for solving a system of linear equations Ax = b is reported. The effect of calibration, the convergence reliability of the BOC, and the convergence of problems with singular matrices are studied.

This paper discusses, for the first time, the application of luminescent rebroadcasting devices to computing. The devices are described, together with some preliminary experimental characterization. The devices are then shown to be capable of: analog addition and multiplication, sequences of ORs and NORs, and performing the 16 Boolean operations possible between two variables.

Historical objections to optical digital computing are critically examined. Theoretical limits for fiber optic fan-in and fan-out factors are discussed. The potential development of highly sophisticated optoelectronic programmable logic devices (OPLA's) based on high fan factor designs is considered. Statistical measurements of a fiber optic fan-out factor of 2940 are presented.