During the fifty years since its initial development as a means of providing early warning of airborne attacks against allied countries during World War II, radar systems have developed to the point of being highly mobile and versatile systems capable of supporting a wide variety of remote sensing applications. Instead of being tied to stationary land-based sites, radar systems have found their way into highly mobile land vehicles as well as into aircraft, missiles, and ships of all sizes. Of all these applications, however, the most exciting revolution has occurred in the airborne platform arena where advanced technology radars can be found in all shapes and sizes...ranging from the large AWACS and Joint STARS long range surveillance and targeting systems to small millimeter wave multi-spectral sensors on smart weapons that can detect and identify their targets through the use of highly sophisticated digital signal processing hardware and software. This paper presents an overview of these radar applications with the emphasis on modern airborne sensors that span the RF spectrum. It will identify and describe the factors that influence the parameters of low frequency and ultra wide band radars designed to penetrate ground and dense foliage environments and locate within them buried mines, enemy armor, and other concealed or camouflaged weapons of war. It will similarly examine the factors that lead to the development of airborne radar systems that support long range extended endurance airborne surveillance platforms designed to detect and precision-located both small high speed airborne threats as well as highly mobile time critical moving and stationary surface vehicles. The mission needs and associated radar design impacts will be contrasted with those of radar systems designed for high maneuverability rapid acquisition tactical strike warfare platforms, and shorter range cued air-to-surface weapons with integral smart radar sensors.

A compact ultra-wideband radar construction is underway, and relevant components are being built or are in the process of development. The initial emphasis of our program is to study the state-of-the-art in EM impulse generators, antennas, receivers, samplers, and digitizers to set a baseline system design for a compact ultra-wideband radar. The components meeting the system design will then be integrated into a functional, self-contained, computer controlled unit. The long term goal of this system is for media identification and media interface detection. Related technology advancement will be described and illustrated in this paper.

The P-3 UHF-UWB system offers a unique capability for remote sensing of forest regions because the low-frequency waveform penetrates foliage and the time resolution allows imaging of individual tree trunks. A tree trunk combined with the ground under it forms a dielectric top hat reflector which is a dominant scattering mechanism. The scatter from the ground-trunk interface is strongly dependent on polarization with Horizontal incidence-Horizontal reflection typically being much stronger than Vertical incidence-Vertical reflections. The cross-polarized components of the scatter is typically much weaker; however, heavily sloped terrain can cause a polarization rotation upon reflection which can cause significant cross-polarization. Estimates of the cross-polarized image can therefore be used to determine terrain characteristics.

A new, fast algorithm for synthetic aperture radar (SAR) image formation is introduced. The algorithm is based on a decomposition of the time domain backprojection technique. It inherits the primary advantages of time domain backprojection: simple motion compensation, simple and spatially unconstrained propagation velocity compensation, and localized processing artifacts. The computational savings are achieved by using a divide-and-conquer strategy of decomposition, and exploiting spatial redundancy in the resulting sub-problems. The decomposition results in a quadtree data structure that is readily parallelizable and requires only limited interprocessor communications. For a SAR with N aperture points and an N by N image area, the algorithm is seen to achieve O(N²logN) complexity. The algorithm allows a direct trade between processing speed and focused image quality.

This paper describes recent performance-enhancing modifications made to the AN/APG-76 radar. An interferometric radar equipped with a four-channel receiver and a seven-channel interferometric antenna, the AN/APG-76 has been used to demonstrate novel interferometric imaging concepts. Originally built as a tactical radar with air-to- air modes, SAR, and three-channel DPCA-like MTI modes, the modified radar's capabilities include: real-time autofocused imaging at 3- and 1-foot resolutions, elevation interferometric SAR (both single and repeat pass), polarimetric imaging, precision tracking by means of a tightly-coupled GPS-aided INS system, and moving target imaging using the inherent clutter-cancellation capabilities of the radar. The re-programmability of the on-board processor allows new real-time modes to be implemented, and high-speed data recording allows off-line analysis of data.

There is an increasing interest in imaging radar systems operating at low frequencies. Examples of military and civilian applications are detection of stealth-designed man- made objects, targets hidden under foliage, biomass estimation, and penetration into glaciers or ground. The developed CARABAS technology is a contribution to this field of low frequency SAR imagery. The used wavelengths offer a potential of penetration below the upper scattering layer in combination with high spatial resolution. The first prototype of the system (CARABAS I) has been tested in environments ranging from rain forests to deserts, collecting a considerably amount of data often in parallel with other SAR sensors. The work on data analysis proceeds and results obtained so far seem promising, especially for application in forested regions. The experiences gained are used in the development of a new upgraded system (CARABAS II), which is near completion and initial airborne radar tests for system verifications followed by some major field campaign are scheduled to take place during 1996. This paper will summarize the CARABAS I system characteristics and system performance evaluation. The major imperfections discovered in the radar functioning will be identified, and we explain some of the modification made in the system design for CARABAS II. A new algorithm for future real-time CARABAS data processing has been derived, with a structure well-suited for a multi-processor environment. Motion compensation and radio frequency interference mitigation are both included in this scheme. Some comments on low frequency SAR operation at UHF-based versus VHF-band will be given.

The primary purpose of GeoSAR is to demonstrate the feasibility of interferometric topographic mapping through foliage penetration. GeoSAR should become a commercially viable instrument after the feasibility demonstration. To satisfy both requirements, we have designed a dual frequency (UHF- and X-band) interferometric radar. For foliage penetration, a lower frequency (UHF) radar is used. To obtain better height accuracy for low backscatter areas, we proposed a high frequency (X-band) interferometric system. In this paper, we present a possible GeoSAR system configuration and associated performance estimation.

Focus of a SAR image during maneuvers requires accurate estimates of the aircraft velocity vector. Velocity errors particularly along the radar line of sight (LOS) to the SAR map center cause LOS acceleration errors during an aircraft maneuver. This LOS acceleration results in defocused SAR images. Sensitivity to acceleration errors is greatest when mapping at small squint angles with respect to the aircraft velocity vector, i.e., at small angles with respect to the aircraft fore-aft axis. This paper shows the expected SAR map focus performance during maneuvers as a function of velocity errors. Simulation results showing expected focus performance using current and advanced processing techniques with background radar Doppler updates and with the use of Global Positioning System (GPS) are presented. With GPS aiding, velocity accuracy is improved by a factor of ten. The simulation results are compared to the flight test results performed with the APG-76 radar using both a gimbaled and a strapdown one nautical mile per hour class Inertial Navigation System with and without GPS aiding. The AN/APG-76 is a SAR radar with simultaneous Ground Moving Target Indication and is designed for `near the nose' imaging during aircraft maneuvers.

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). The AO processor will replace the function of several digital processor boards currently in the radar. The primary objective of this program is the real-time demonstration of an optical processor in the MICOM radar. This paper provides an overview of the MICOM radar system, discusses the design of the AO range-Doppler processor, and describes the radio frequency and digital electronic interfaces required to achieve real-time operation in the MICOM radar. Upcoming integration and test activities are then described.

This paper provides radar test results of an optical processor for sidelobe jamming reduction. Using a Rome Laboratory C-Band phase array radar and jammer testbed, numerous realistic test scenarios were performed. These tests were conducted with the radar in a receive-only mode, with target and jammer signals provided externally. The sum beam from the C-band array served as the main channel input to the adaptive canceller, and one of eight subarrays provided the auxiliary channel. Targets and jammers consisted of both barrage noise and pulsed continuous tone signals produced by stationary jamming and target sources located in the far field of the phase array radar. Closed loop testing, sidelobe jamming reduction, and multipath signal considerations were all integral parts of the test scenarios. The processor is designed to cancel 10 MHz wide bandwidth jamming signals and to work at the radar intermediate frequency of 80 MHz. A processor overview, cancellation results for a variety of test conditions, and future hardware enhancements and test plans will be summarized.

IF (intermediate frequency) sampling is a method of sampling the received radar waveform out of the IF channel directly, without mixing to baseband, using a single A/D converter. The sampling rate needed is a multiple of the bandwidth of the IF filter, of the order of 3 times the -3 dB bandwidth. IF filter skirt attenuation limits aliasing effects and permits apparent undersampling of the IF frequency. Stretch processing is the method of matching the radar's LO frequency ramp rate (linear FM) to the transmit waveform's `chirp', in order to limit the IF bandwidth requirement to a value much less than the RF bandwidth and thus permit a lower rate of sampling. The combination of IF sampling and stretch processing is advantageous because A/D samplers are now able to operate at adequately short sample- and-hold aperture times, for use at IF frequencies, with a good number of bits resolution, and stretch processing can use narrow IF bandwidths. Therefore, high range resolution can be achieved at a lower cost than with quadrature channels at baseband and dual A/D's. Added benefits are the elimination of I-Q imbalance effects, A/D DC offset effects, and the need for calibration of these effects. Some A/D saturation can also be tolerated. A Fast Fourier Transform of the real sample data set is easily converted to an inphase and quadrature output data set for further operations. The paper goes into the equations and methodology of such a radar system and delineates the hardware differences between the baseband approach and the IF sampling approach.

A radar interferometric topography mapper designed to acquire digital elevation maps of the earth's surface from the Space Shuttle is described and its performance estimated. The system described is capable of acquiring a topographic map of all of the earth between 54 degree(s)S and 60 degree(s)N latitude to a height accuracy of 16 meters absolute. This planned mission will be the first use of radar interferometry to acquire topographic data on a global scale, the data of which will have significant impact of many applications. The system uses the previously flown SIR- C C-Band synthetic aperture radar system augmented by a second interferometric antenna deployed 60 meters from the Shuttle. The operation of the system, which requires the simultaneous use of dual polarization radars operating with horizontal and vertical polarizations with electronic beam scanning, is described. Performance parameters which drive the vertical height accuracy of this system and the implementation of solutions necessary to meet the performance objectives are described.

An analysis of a digital/adaptive beamforming architecture consisting of a main array, and 1 to n shared auxiliary arrays formed by digital beamforming among a group of shared elements was conducted to determine the effectiveness of jamming interference suppression in the sidelobes of the main array when the jammer angle of arrival is known `a priori'. The auxiliary array outputs are combined with the main array output to cancel incoming jamming signals in the sidelobes of the main antenna. A single shared element auxiliary array can be used to cancel multiple jammers performing digital beamforming among the auxiliary array elements. With the use of peripheral sensors such as an ESM receiver, jammer angle of arrival information is used to steer the main antennas in the direction of a known target signal, and the auxiliary array in the direction of one (or more) known jamming signals. By digitally beamsteering the auxiliary array in the jammer directions, the sidelobe degradation in the main array is minimized. After steering the main and auxiliary arrays, complex weights are determined. The presence of a desired target signal in the mainbeam, along with the jamming signals in the sidelobes can result in weight contamination. This problem is avoided by adapting to the jamming in a pre-listen, receive-only mode before a radar dwell is transmitted. Sidelobe cancellation performance analysis results will be presented for the cases of narrowband and wideband steady-state jamming.

The use of pulse compression waveforms such as binary phase codes and frequency coded signals enables a high level of resolution required for Earth-based planetary mapping. The ambiguity function is a useful tool that reveals the resolution that can be attained when using a specific waveform. We derived a general expression for the ambiguity function for these random signals to show the resolution for both waveforms. In addition, the expected level of signal self-noise for binary phase codes was determined. The resolutions in delay and Doppler are obtained from the expressions. The resulting plots of derived expected values of ambiguity functions for the random waveforms are given. In addition, plots of the derived expressions for the mean square and variance of the ambiguity functions are also shown for binary phase codes. These show sidelobe behavior (signal self-noise) in delay and Doppler. The square of the mean of the ambiguity functions for each waveform is then convolved with a Gaussian scattering function that models the overspread characteristics of the planet Mars to demonstrate the comparably similar resulting images with negligible spreading. The system parameters required to obtain the images with specified resolutions are compared to show the advantages or disadvantages of using each waveform.

Over the past several years Northrop Grumman has been developing Non-Cooperative Target Recognition (NCTR) technology using High Range Resolution (HRR) Radar data. Common to all NCTR efforts is the need to train classifier algorithms on limited sources of data. The classifier design must also address signature variations with aspect viewing angles and stores configurations. This paper will provide a methodology for quantifying training data segmentation issues including: (1) Degradation due to limited samples within an aspect zone; (2) Stability of scattering centers as a function of aspect angle; and (3) Stores variations. In a program supported by Wright Patterson AFB, Northrop Grumman has developed a detailed statistical model of the Airborne Radar Target Identification HRR signature data. The statistical model is based on a template alignment procedure. This model provides an analytic basis for predicting classifier performance using an associated distance metric. This paper will provide a brief discussion of our template classifier and apply the analytic model to the segmentation issues in the previous paragraph.

The faithful simulation of a semi-active missile is a very complex system encompassing numerous technical issues. The simulation problem is complicated because part of the simulation, the radar illuminator, is earth based while the radar receiver is in motion and the simulation must have six degrees of freedom. The simulation issues are examined and solutions provided, based on three decades of simulator development experience of Georgia Tech.

The Georgia Tech Research Institute, sponsored by the Warner Robins Air Logistics Center, has developed an approach for efficiently postulating and evaluating methods for extending the life of radars and other avionics systems. The technique identified specific assemblies for potential replacement and evaluates the system level impact, including performance, reliability and life-cycle cost of each action. The initial impetus for this research was the increasing obsolescence of integrated circuits contained in the AN/APG-63 system. The operational life of military electronics is typically in excess of twenty years, which encompasses several generations of IC technology. GTRI has developed a systems approach to inserting modern technology components into older systems based upon identification of those functions which limit the system's performance or reliability and which are cost drivers. The presentation will discuss the above methodology and a technique for evaluating and ranking the different potential system upgrade options.

The increasing speeds of modern-day threats necessitate interceptor platforms that must generate matching capabilities in velocity and maneuverability. The interceptor's size and weight limits the available antenna aperture, and consequently, the broad antenna beam that results combines with the high platform speed to smear returning echoes over a larger frequency spectrum than is encountered in more traditional airborne radars. Furthermore, the radar must perform in very high-g environments that are typical of such engagements. The very high closing speeds dictate short engagement times that in turn limit the time available for the radar to acquire a target and then track it with acceptable accuracy. We describe a radar design that attempts to reconcile these (at times, conflicting) requirements.

The Army Research Laboratory is evaluating ultra-wide-band radar imaging techniques for subsurface target detection. Of importance in this effort is the generation of an appropriate waveform and the development of an ultra-wide- band exciter (UWBE). A critical requirement for ground penetrating (GPEN) radar is to identify near-surface or subsurface targets in sufficient detail to allow unambiguous identification. For example, a subsurface mine must be distinguishable from benign subsurface soil strata or other man-made objects (e.g., decoys). To optimize the measured signal-to-clutter ratio requires high cross-range and down- range resolution. High cross-range resolution is achieved by collecting radar returns while the radar is in motion and using synthetic aperture techniques to process those returns. High down-range resolution is achieved by the transmission of wide-bandwidth waveforms. The UWBE design uses a wide-bandwidth (approximately 3 GHz) linear frequency modulated (LFM) waveform. Another critical requirement for GPEN radar is the need for efficient propagation of the radar waveform into the soil, which enhances the detection and recognition of subsurface objects. Since low frequencies (approximately 10 MHz) propagate better into soils than do high frequencies, a low chirp start frequency is desired. The use of a LFM waveform that spans from HF to S Band presents another problem of co-site interference with commercial communication equipment (FM, TV, and cellular radio). Since broadcasting in these bands is restricted, a method has been developed to arbitrarily notch out portions of the transmitting bandwidth. This paper will discuss the use of an arbitrary waveform generator (AWG) from Tektronix with a switching local oscillator (LO) architecture to generate the low start frequency wideband LFM waveform required. The AWG with its 1 GHz clock is bandwidth limited to approximately 400 MHz by Nyquist sampling and filter design constraints. The longer LFM waveform is generated by frequency offsetting and concatenating multiple LFM waveform packets from the AWG. The frequency offset is controlled by the switching LO architecture, where the switching time is on the order of a few nanoseconds. Each AWG output can be pre-programmed with notches in the band for interference suppression, as well as a phase offset to maximize the phase continuity of the desired LFM waveform.

The Army Research Laboratory (ARL) has been investigating the potential of ultra-wideband synthetic aperture radar (UWB SAR) technology to detect and classify targets concealed by subsurface targets and foliage. Our investigative approach is to collect high-quality precision data to support phenomenological investigations of electromagnetic wave propagation through dielectric media. These investigations, in turn, support the development of algorithms for automatic target recognition. In order to achieve these goals, ARL designed and built an impulse (very short pulse) radar to collect data at a variety of test sites to measure and analyze the responses from targets, clutter, and targets embedded in clutter. The UWB BoomSAR, mounted on a 150-foot-high mobile boom lift, collects the high-quality, precision data sets needed for understanding UWB SAR system requirements and foliage penetration and ground penetration phenomenology. The BoomSAR operates with over 1 gigahertz of bandwidth covering a spectrum from 40 MHz to 1 GHz and is fully polarimetric. This bandwidth contains low frequencies needed for ground penetration while also maintaining higher frequency coverage for high resolution imagery. This paper shows a GPEN target area from data collected at Yuma Proving Grounds, AZ in low- and high- frequency subbands.

A 2D finite-difference time-domain (FDTD) algorithm is presented for modeling a canonical target (disk) buried in realistic soil at Yuma Proving Grounds. The buried disk was measured by a ground-penetrating radar. Scattering phenomena observed in the FDTD modeling results are detailed and sampled fields are compared with measurements.

The Method of Moments is utilized to compute the complex resonant frequencies and modal currents of perfectly conducting wires and bodies of revolution buried in a lossy, dispersive half space. To make such an analysis tractable computationally, the half-space Green's function is computed via the method of complex images, with appropriate modifications made to account for the complex frequencies characteristic of resonant modes. Results are presented for wires and bodies of revolution buried in lossy soil using frequency-dependent measured parameters for the complex permittivity, and we demonstrate that the resonant frequencies generally vary with target depth.

The method of moments is used to analyze short-pulse plane- wave scattering from perfectly conducting thin wires and bodies of revolution buried in a lossy, dispersive half space. The analysis is performed in the frequency domain, with the time-domain fields synthesized via Fourier transform. To make this analysis efficient, the method of complex images is used to compute the frequency-dependent components of the half-space dyadic Green's function. Results are presented for short-pulse scattering from buried wires, spheres and cylinders, using measured frequency- dependent soil parameters (permittivity and conductivity), and the phenomenology associated with the scattering is investigated in detail.

The finite-difference time-domain (FDTD) algorithm, when modified to include dispersion, is a convenient tool for analyzing wave propagation in lossy, dispersive soils. It has been shown recently that a convenient way to include the dispersion is through a spatial array of digital permittivity or conductivity filters, one filter at each space node in the finite-difference grid. We address here the problem of selecting filter coefficients that result in accurate and stable FDTD implementations. We present a systematic procedure for determining the coefficients based on a nonlinear optimization procedure and stability test.

The Focused Array Radar (FAR) is a unique time-domain radar system which uses adjustable time delayed signals in a wide multi-element array which focuses transmitted and received signals to detect targets in lossy soil. By making use of specially-designed folded rhombus antenna elements--which are both ultra-wideband and more omnidirectional in the forward direction than a comparable dipole--the FAR optimizes the trade-off between target resolution and penetration depth. These projectory antenna elements, patented by GEO-CENTERS, INC., faithfully radiate sub- nanosecond pulses with frequency response varying from about 700 MHz to 1.3 GHz, so targets in wet soils within 60 cm of the surface and as small as 8 cm can be resolved. The array signals are focused by establishing the time delay from each element to each sample point in the soil medium, taking into consideration the differing propagation speed in air and various soils, as well as ray path refraction at the air/ground interface. These delays are applied in roughly ten picosecond intervals to the transmitted signal and used to time gate the received signal. By using time delays for a focused wideband pulse, the phases of each frequency component of the radar signal are in effect properly specified for constructive interference. Also, as a result of the time-gating, the large ground surface reflection signal is avoided.