An optical associative memory provides rapid content addressing of freeform alphanumeric text stored on an optical disk. A highly parallel, space-integrating intensity correlator has been breadboarded using off-the-shelf components for proof of concept and identification of component characteristics. Performance is found to be sensitive to light source coherence, optics modulation transfer function, media bit placement and contrast uniformity. Simplified processing algorithms extendable to on-chip implementation are evaluated with simulated and experimental results. Conventional codes cannot be optimized to provide a unique coding for a given work, nor do they have a constant optical return from large-area illumination; these are essential to facilitate the correlation readout. A custom code has been developed for correlation readout optimization.

The TOPS pattern recognition program produced a successful demonstration of optical correlation applied to automatic target recognition (ATR) using visible (video) images. The TOPS correlator hardware will soon be tested performing ATR using IR images. In both cases, effective algorithms must be defined and implemented in three functional areas: (1) distortion-tolerant correlation filter design, (2) input image preprocessing, and (3) filter management strategy. The algorithms implemented for TOPS are reviewed and the modifications required for IR imagery are discussed.

Martin Marietta has successfully completed a TOPS optical pattern recognition program. The program culminated in August 1994 with an automatic target recognition flight demonstration inwhich an M60A2 tank was acquired, identified, and tracked with a visible seeker from a UH-1 helicopter flying a fiber optic guided missile (FOG-M) mission profile. The flight demonstration was conducted by the US Army Missile Command (MICOM) and supported by Martin Marietta. The pattern recognition system performance for acquiring and identifying the M60A2 tank, which was positioned among an array with five other vehicle types, was 90% probability of correct identification and a 4% false identification for over 40,000 frames of imagery processed. Imagery was processed at a 15 Hz input rate with a 1 ft³, 76 W, 4 GFLOP processor performing up to 800 correlations per second.

Radio frequency signal processing systems that use photorefractive crystals to provide key functions are described. Some new experimental results obtained with a time-integrating correlator developed to support the acousto-optic null steering processor effort are presented that demonstrate the effectiveness of the photorefractive time-integrating approach. We also describe a new adaptive notch filtering system that provides effective and flexible excision of narrowband interference. The system architecture makes optimal use of the unique properties of the photorefractive effect to provide a simple yet highly effective solution for the adaptive signal processing problem.

The development of an acousto-optic null steering processor for radar ECCM application is reviewed. The general problem to be solved by this special purpose signal processor is discussed and the advantages of the acousto-optic approach are given. Processor architecture, AO/EO component selection, and experimental discoveries are described. Measurements of the laboratory breadboard model's performance for a variety of multipath and non-multipath signals are also covered. Limitations of the current breadboard model and future directions are discussed.

We report our progress on integrated phosphorus-doped SiO₂ optical waveguide delay lines and membrane optical routing switches for phased-array radar control. We have completed the design and layout of the delay lines for the two shortest bits. We have demonstrated the concept of a microelectromechanical membrane optical routing switch with a Mach-Zehnder interferometer and a fixed aluminum thin film. Channel crosstalk values as low as -12.4 dB and -20.3 dB were measured with and without a 3 mm aluminum film, respectively. We have designed the membrane structure for the switch to have better yield, improved reliability, and lower excitation voltage.

An L-band 96-element array controlled by photonics was developed and tested. A time shift beamforming network provides a 50% instantaneous bandwidth (850 MHz-1400 MHz). The antenna beam was scanned to +/- 60$DEG without beam squint over the entire bandwidth. A nanosecond impulse response has been measured to demonstrate a range resolution of better than 30 cm for target ID and imaging.

A wideband time-domain approach is described for the calibration of an L-band photonic array antenna using time shifters. A nanosecond pulse was injected into each channel with the time shifter cycling through all the states so that the insertion loss and the time delays could be calibrated. The pulse was not generated in real time; it was instead synthesized with 801 frequencies over a wideband (50%) from 850 to 1400 MHz. The network analyzer transmits the cw frequencies one at a time over a period of 100 ms, and the measured data are transformed into the time domain with the built-in FFT.

The modern threat for electronic support measure (ESM) receivers are radars that use exotic wideband modulation waveforms for achieving higher resolution or reducing the probability of intercept. Existing ESM testbeds can receive both narrowband and wide-bandwidth modulations but can only detect the presence of narrowband signals. A wideband acousto-optic (AO) correlator (AOC) has been jointly developed by the Army Research Laboratory and Dynetics, Inc., under the ARPA TOPS program for insertion into an existing ESM testbed. The AO module is an in-line, time-integrating correlator architecture that offers a small, lightweight solution for detecting and analyzing inherently wide-bandwidth, spread-spectrum signals. The correlator processes 500 MHz of instantaneous bandwidth and offers enhanced detection capability of direct-sequence, phase-modulated chirps, and frequency-hopping signals. The ESM testbed, developed by the Intelligence and Electronic Warfare Directorate of the US Army Communications and Electronics Command, is currently being evaluated for integration into existing and future Army platforms. The AOC complements the existing ESM testbed and offers a wideband detection capability as described in this paper. The AOC insertion with the ESM testbed was demonstrated at the International Low Probability of Intercept Trials in Salisbury, australia. Performance results for the AOC against realistic LPI waveforms from this field test will be presented.

Miniaturization and ruggedization is paramount in the application of optical signal processing systems to modern electronic combat (EC) applications. This paper describes two state-of-the- art fabrication methods used for miniaturizing acousto-optic processing modules.

An improved optical technique for electrical phase measurement has been developed at the Naval Research Laboratory under ARPA sponsorship. The approach has been demonstrated in an rf interferometer application and its performance evaluated under condition of direct rf injection and free-space irradiation. Test results indicate high-accuracy angle of arrival and low probability of ambiguity was achieved. This paper describes the phase measurement concept, implementation into the NRL Precision Direction Finding Receiver brassboard unit, calibration techniques, automated test software, and measurement results.

Passive direction finding is an integral function of EW and ELINT systems. The ability to produce direction-of-arrival information in a timely, accurate manner is strongly influenced by both the direction-finding techniques employed and the processing algorithms. The processing algorithms must detect a signal, classify it as being a signal of interest, extract the signal's parameters, including its direction of arrival, control the receiver, and interface with the system's mission computer. This paper describes a direction-finding receiver that used an optical processor to extract the direction-of-arrival information from the signal environment and the processing algorithms required to support the optical processor.

The EW channelizer and PDF were developed under the ARPA TOPS program to satisfy broad electronic warfare systems demands. This paper describes EW system architectures and the functions the TOPS EW channelizer and PDF processor performs in these systems. EW system architectures examined includes Ships Early Warning, airborne Surveillance, Weapons, and Airborne Warning. In addition, functional descriptions of the EW and PDF processors are provided.

Progress made during the previous 12 months toward the fabrication and test of a flight demonstration prototype of the acousto-optic time- and space-integrating real-time SAR image formation processor is reported. Compact, rugged, and low-power analog optical signal processing techniques are used for the most computationally taxing portions of the SAR imaging problem to overcome the size and power consumption limitations of electronic approaches. Flexibility and performance are maintained by the use of digital electronics for the critical low-complexity filter generation and output image processing functions. The results reported for this year include tests of a laboratory version of the RAPID SAR concept on phase history data generated from real SAR high-resolution imagery; a description of the new compact 2D acousto-optic scanner that has a 2D space bandwidth product approaching 10⁶ sports, specified and procured for NEOS Technologies during the last year; and a design and layout of the optical module portion of the flight-worthy prototype.

We describe some performance trades for a hybrid optical processor for real-time synthetic aperture radar (SAR) image formation. A 2D Fourier transforming time-integrating interferometrically based optical processor is a key element of the system. The optical processor uses a modulated laser diode for radar signal insertion, crossed 1D acousto-optic scanners for 2D scanning, a modified Koster interferometer for fringe generation, and fast detector arrays (cameras) for light detection and integration. The image dynamic range is enhanced by processing many camera frames. Digital pre- and post-processing play essential roles in the system enhancement. We present the characteristics of this type of processor and consider some of its performance trades. The optical processor design approach lends itself to the important attributes of high (real-time) data rates, multiple SAR mode processing capabilities, compact and rugged packaging, and power efficiency.

The primary objective of this program is to perform pulse compression in an existing military system using an optical processor. To achieve this objective, an acousto-optic (AO) range- Doppler processor is being developed to interface to an advanced ground-based radar system developed by the U.S. Army Missile Command (MICOM) and replace the function of several digital processor boards currently in the radar. This paper provides an overview of the MICOM radar system, describes the AO range-Doppler processor, the rf and digital electronic interfaces required to achieve real-time operation in the MICOM radar, and system integration issues.

The interfacing and post-processing requirements for the development and insertion of an acousto-optic (AO), range-Doppler processor will be described. This system has been configured to operate as an integral part of the signal processing chain of an advanced spread- spectrum radar developed by the US Army Missile Command (MICOM). This MICOM radar transmits a continuous repeated, biphase-coded waveform and processes a block of received data to detect and track targets i range and Doppler in the presence of severe ground clutter. Multiple code rates are processed to extend the range window through application of residue number techniques. Range and Doppler processing are achieved in the AO processor using an additive triple-product processor architecture that coherently detects the range-Doppler information on a high-speed, custom 3D CCD detector array developed by the Army Research Laboratory. We present the interfaces to the radar and the post-processing of the data produced by the AO range-Doppler processor into the format required by the MICOM signal processor. The interfaces comprise the extraction of digital in-phase and quadrature data, the condition of this data for the AO range-Doppler processor, and the insertion of the post- processed optical data into the radar signal processor. Timing and latency issues are critical to real-time operation (creating range-Doppler images at approximately 1600 Hz frames rates) within the MICOM radar. The post-processing section cover optical processor architecture/post-processing tradeoffs, focusing on requirements, algorithms, and hardware implementation.