Recent advances in millimeter-wave (MMW) radar technologies provide new applications for law enforcement use over-and- above the venerable speed timing radar. These applications include the potential to detect weapons under clothing and to conduct surveillance through walls. Concealed Weapon Detection and covert surveillance are of high interest to both the Department of Defense in support of Small Unit Operations and the Justice Department for civilian law enforcement applications. MMW sensors are under development which should provide the needed capabilities including radiometric sensors at 95 GHz, active 95 GHz real aperture radars, active focal plane array (FPA) radars, and holographic radars. Radiometric sensors include 2D FPA systems, 1D FPA, scanned systems, and single element scanned sensors. Active FPA radars include illuminated radiometric systems and coherent radar systems. Real aperture MMW radar systems include raster scanned and conical scanned sensors. Holographic systems ruse mechanical scanners to collect coherent data over a significant solid angular sector.

Ground penetrating radar (GPR) is becoming an increasingly useful tool for road subsurface characterization. The Florida Department of Transportation (FDOT) has recently obtained a new 1 GHz ground penetrating radar with the ability to make high resolution measurements. Depth profile scan rates of the new radar are about 50 scan/sec and the radar operates on a test van travelling at speeds up to 50 - 55 MHz. The time domain data collected by the GPR allow the determination of thickness of the road surface and subsurface layers and, with appropriate signal processing, the data can provide information about voids and other anomalies within road layer interfaces. This paper will describe the salient features of the Florida DOT ground penetrating radar, recent measurement results, and applications of GPR for road assessments. It will also describe preliminary results of a University of Florida project which is employing advanced signal processing techniques to detect and classify subsurface anomalies in road layers. As a precursor to anomaly detection we are developing improved techniques for finding road layer thicknesses and dielectric constants. The processing techniques being developed include matched filter and slope detection algorithms. A goal of the current work is to develop signal processing techniques that will allow FDOT to evaluate subsurface conditions for large sections of road throughout Florida in a more accurate and rapid manner. It is expected that the GPR and the results of current research will assist the FDOT in more accurately determining road layer thickness profiles, in assessing road subsurface conditions with less coring, and aid in rehabilitating roads with less manpower than is now required. Such capabilities will allow potentially serious problems to be corrected before they become costly and will also provide a useful tool for future road design and improvement.

Researchers at the Georgia Tech Research Institute have developed a radar that will detect heartbeat and respiration without any physical connection to the subject. The system is capable of making these measurements at ranges exceeding 10 meters. This paper explores the use of the system for the biometric identification of personnel who work in a highly secure environment. The system, used in this application, would use the heartbeat signature of an individual as a biometric identifier. Also, the system could be used to determine the stress level being experienced by an individual on the basis of respiration and heartbeat rates.

Currently available radar instruments are not capable of guiding a helicopter pilot safely during approach and landing under poor visibility conditions. This is due to lack of resolution and lack of elevation information. The RADAR technology that promises to improve this situation is called ROSAR, which stands for Synthetic Aperture Radar based on ROtating Antennas. In 1992 Eurocopter and Daimler- Benz Aerospace investigated the feasibility of an imaging radar based on ROSAR technology. The objective was to provide a video-like image with a resolution good enough to safely guide a helicopter pilot under poor visibility conditions. ROSAR proved to be especially well suited for this type of application since it allows for a stationary carrier platform: Rotating arms with antennas integrated into their tips can be mounted on top of the rotor head. In this way the scanning region of the antennas can cover 360 degree(s). While rotating, the antenna scans the environment from various visual angles without assuming a movement of the carrier platform itself. The signal is then processed as a function of the rotation angle of the antenna movement along a circular path. A radar system of this type is now under development at Eurocopter and Daimler-Benz Aerospace: HeliRadar. HeliRadar is designed as a frequency modulated continuous wave radar working in a frequency band around 35 GHz. The complete transmitter/receiver system is fixed mounted on top of the rotating axis of the helicopter. The received signals are transferred through the center of the rotor axis down into the cabin of the helicopter, where they are processed in a high performance digital signal processor (processing power: 10 GFLOPS). First encouraging results have been obtained from an experiment with `slow motion' movement of the antenna arm.

Inverse synthetic aperture radar (ISAR) imaging on a turntable-tower test range permits convenient generation of high resolution 2- and 3-D images of radar targets under controlled conditions, typically for characterization of the radar cross section of targets or for testing SAR image processing and automatic target recognition algorithms. However, turntable ISAR images suffer geometric distortions and zero-Doppler clutter (ZDC) artifacts not found in airborne SAR images. In this paper, ISAR images formed at Georgia Tech's Electromagnetic Test Facility are used to demonstrate and compare selected members of one family of 2- D ISAR imaging algorithms, from a simple but distortion- prone 2D discrete Fourier transform to a computationally- intensive matched filter solution. A simple algorithm for correcting range curvature using image domain resampling is described. We then demonstrate two signal processing techniques to suppress zero-Doppler clutter while minimizing effects on the target signature. The first removes ZDC components in the frequency domain, whereas the second performs cancellation in the image domain.

In this work the problem of developing a theory of sea scatter for ultrawideband electromagnetic waves is examined. The motivation for this work derives from previous work of Wetzel, who formulated a scalar, time-domain model of sea clutter and showed that the local curvature properties of the rough surface are an important factor in determining the sources of the scattered return. Here, Wetzel's result is extended to the electromagnetic problem and then extended further to obtain some polarization effects consistent with the small perturbation method.

Due to recent development of the theory related to non- uniform Pulse Repetition Intervals (PRI) for pulse doppler radars, it has become feasible to combine the multiple pulse integration of Synthetic Aperture Radar (SAR) with the temporal agility of non-uniform PRI's. This combination can bring to the traditional SAR system the benefits of both strong resistance to countermeasures and interception, and high sensitivity. A significant problem which must be addressed in synthetic aperture radars is the possible smearing of the azimuth dimension of the image due to doppler sidelobes which are too high. This issue is even more serious with a non-uniform PRI waveform. In a traditional SAR the sidelobes are controlled with signal processing techniques such as amplitude weighting. However, these simpler techniques are not adequate for non-uniform PRI waveforms. One method for increasing the sidelobe rejection capability of these waveforms requires the use of a processor which consists of time varying linear digital filters. One way to design such a filter is through the use of interpolation based techniques.

This paper explores the use of nonlinear prediction in the modeling of sea clutter. The nonlinear methods considered here are based on local approximations: nearest neighbor and local linear prediction. The effects of radar scan ranges and the number of training samples on the nonlinear clutter model are examined. The nonlinear predictive model is then used for clutter suppression to enhance target detectability. It is shown that the nonlinear predictive detection scheme can detect small floating targets such as beach balls embedded in sea clutter. The standard linear prediction is used for comparison. It is observed that the nonlinear prediction outperforms the linear one on a regular basis.

The objective of the SAIP ACTD is to make imagery a responsive contributing source to a commander's overall battlespace awareness by focusing on theater and tactical sensor exploitation. The goal of the exploitation system is to increase the image analyst efficiency in exploiting large volumes of image data produced by current and future theater and tactical imaging platforms. The system will be evaluated based on its ability to improve the analyst's capability to detect and recognize isolated targets, minimize false alarms, recognize force structure (e.g. maneuver battalions), and provide a closed loop cueing of spot mode from strip imagery.

MIT Lincoln Laboratory has developed a complete, end-to-end, automatic target detection/recognition system for synthetic aperture radar data. The system uses resolution enhancement (super-resolution) techniques to improve the performance of the automatic target recognition stage. This paper presents a new multi-resolution classification scheme that greatly improves the computational efficiency of the classifier with only a slight loss in classification performance.

In this paper, we introduce the 2D continuous wavelet transform (CWT) as a tool for the detection stages of an SAR ATR system. We demonstrate that the 2D CWT tuned to reasonable target sizes can enhance the signal to clutter radio and improve detection performance in a focus of attention algorithm as compared to the traditional CFAR algorithm. We also show that the 2D CWT can be used to estimate the size and pose of a target, which can provide important features for second level detection. The detection and feature extraction algorithms were tested on measured one foot resolution SAR data.

In this paper we present an analysis of distance transform methods of matching object models to SAR data. We show that by properly defining the distance function, the likelihood of each observed SAR feature data point given the model is given as a function of position. This allows calculation of a likelihood of observing a set of data features, given a model and its associated pose and other parameters. The issue of normalization resulting from the non-correspondence based distance transform method is discussed. When prior densities of the model features are available, maximum a- posteriori results are obtainable. This method allows the use of priors of models and individual features, along with the geometric probability densities associated with the feature prediction and measurement processes, to be incorporated within a fast correlation-type distance transform matching module. The method also potentially allows exploitation of persistent scatterers over a limited range of SAR model-to-target imaging parameters.

This paper presents an ATR design paradigm that self configures and adapts to the diverse scenarios encountered during a mission. Today's ATR is constructed via inefficient and sub-optimal system configuration and training, whose process is very labor intensive, subjective and inaccurate. The resulting ATR is only capable of a limited amount of adaptation to changes in the environment. Moreover, the operation of such ATR systems require a user with expert algorithmic knowledge. Addressing the above-mentioned problems, the Honeywell effort is producing a self-adaptive ATR system. The system employs a Genetic Algorithm to autonomously and optimally perform configuration and training; the system also includes a specific knowledge capture mechanism, the Context Capture tool, which ties the context of the mission with an optimal configuration. Lastly, the system employs Case Based Reasoning to dynamically configure and control the ATR system based on the changing context during an ATR mission.

One nonlinear adaptive approach to generating target recognition algorithms, the distributed connectionist approach, is also referred to as a neural network. These algorithms frequently employ a gradient descent technique, such as the back propagation learning algorithm, to find a mapping that separates the n dimensional feature space into m recognizable classes. Gradient descent techniques are known to be limited by a characteristic referred to as the `local minima' problem. During the search for an optimum solution or global minima, these techniques can encounter local minima from which they cannot escape due to the `steepest descent' nature of the approach. However, several training techniques used to speed up training or to otherwise optimize these adaptive learning algorithms have side effects which can obviate this local minima problem. We will define a local minima problem with respect to the 1D target ID problem. Appropriate terminology and an error space relevant to the 1D range profile problem will be presented. Four techniques, dynamic architecture definition, weight pruning, adaptive learning rate selection and dynamic training set generation used to optimize training for the multilayer perceptron will be summarized. An analytical explanation of a common underlying mechanism which allows escape from local minima and is shared by these techniques is presented. Some additional advantages are provided by one of the four techniques, the dynamic training set technique. Evidence of these advantages, consistently high quality results, the automatic identification of anomalous signatures in the data base and simple implementation, will be presented.

The previous work demonstrate that detection based on the Laguerre transform exploits the 1D resonance response of the metallic target in the ultra wideband (UWB) synthetic aperture radar (SAR) images. The success is mainly due to the fact that Laguerre function space captures the information contained in the resonant response. However, in that case, only 1D resonance response information were utilized. In this paper, a new CFAR statistics based on spatial resonance templates is proposed. Based on the GLRT, the new CFAR statistics is formulated using the spatial extent of targets' resonance responses in the UWB SAR scenario. In the end, we show that the proposed CFAR statistics yields more robust detection.

Segmentation and labeling algorithms for foliage penetrating (FOPEN) ultra-wideband Synthetic Aperture Radar (UWB SAR) images are critical components in providing local context in automatic target recognition algorithms. We develop a statistical estimation-theoretic approach to segmenting and labeling the FOPEN images into foliage and non-foliage regions. The labeled maps enable the use of region-adaptive detectors, such as a constant false-alarm rate detector with region-dependent parameters. Segmentation of the images is achieved by performing a maximum a posteriori (MAP) estimate of the pixel labels. By modeling the conditional distribution with a Symmetric Alpha-Stable density and assuming a Markov random field model for the pixel labels, the resulting posterior probability density function is maximized by using simulated annealing to yield the MAP estimate.

A method to automatically detect targets from sets of pixel- registered visual, thermal, and range images is outlined. It uses operations specifically designed to work on the different kinds of images to explore the information given by each of them. Five features are used to distinguish the targets from the clutter; texture, brightness, temperature, surface planarity, and height. We describe two different schemes. The first one involves logically combining the results of individual detectors with binary detection information as output. A morphological operation called `erosion of strength n' is introduced and utilized as a powerful tool for removal of spurious information. The second scheme utilizes the concept of `images of interest' to combine the information provided by the different images. A simple linear combination of these images yields excellent detection results. The success of this scheme supports its suitability for other ATR (Automatic Target Recognition) problems.

Based on the statistics theory and the pulse compression technique, a statistical method of reducing the range sidelobe (RSL) of the random binary phase codes (RBPC) is presented, which is different from those of decreasing the RSL of the pseudorandom binary phase codes. The theoretical analysis and computer simulation show that it is possible to suppress the peak RSL to lower than -30 dB, which can effectively guarantee the RBPC radar with good electronic counter-countermeasures feature applicable. Additionally, owing to the Doppler of the target, the maximum loss of the ratio of the mainlobe and sidelobe (MSR) is also discussed. In the meantime, the approach to realization of the RSL reduction with digital signal processors is given.

This paper studies the relationship between high range resolution and signal bandwidth. In order to generate big bandwidth, a novel spread spectrum radar waveform is present in this paper. This paper derives the ambiguous function of intrapulse and interpulse FM radar waveform. The result proves that this kind of waveform stimulus is of big time duration and big bandwidth. This waveform trades off the drawback of interpulse stepped frequency waveform in range resolution. There is not range ambiguous problem in time domain. According to the distinguishing feature of this kind of waveform, a novel two-stage pulse compression system using surface acoustic wave (SAW) device is put forward. It aims to generate a intrapulse rectangular spectrum by means of SAW device. Then, the system can obtain spread rectangular spectrum using interpulse frequency shifting. After matched pulse compression processing, this waveform holds high resolution in range domain and in velocity domain. The system total compression ratio can be controlled by the intrapulse compression ratio and interpulse compression ratio.

The possibilities of radar observation of effects occurring during aerodynamic object flights with sonic and supersonic speeds are considered. Three mechanisms resulting in electromagnetic wave backscattering are analyzed: reflections from shock wave taking place for atmosphere object flight with supersonic speed, reflections from sonic field leading to atmosphere parameter modulation, and scattering from atmosphere turbulence and parameters of which are varied by sonic field influence. The results of experimental study of sonic interaction with atmosphere turbulence are presented. It is shown that the most perspective mechanisms for radar observation are the first and the third ones, the directions of further investigations are put forward.

Radar sensing of multi-layer structures deserves particular attention in the respect of practical applications in agriculture, geophysics, etc. Subsurface radar means allow to increase the value of geologic surveillance. Transferring from sensing homogeneous layers with small losses to sensing structures with great attenuation of signal demands lowering of the frequency of emitted signal. The article considers a method of sensing underground waters, snow and ice layers with frequencies of 20...150 MHz. Coherent and incoherent components of backscattered RF signal's power and envelope variation ratio are regarded the sources of information about the thickness of a layer.

We propose an optimal radar pulse compression technique and evaluate its performance in the presence of Doppler shift. The traditional pulse compression using Barker code increases the signal strength by transmitting a Barker coded long pulse. The received signal is then processed by an appropriate correlation processing. This Barker code radar pulse compression enhances the detection sensitivity while maintaining the range resolution of a single chip of the Barker coded long pulse. But unfortunately, the technique suffers from the addition of range sidelobes which sometimes will mask weak targets in the vicinity of larger targets. Our proposed optimal algorithm completely eliminates the sidelobes at the cost of additional processing.

The Army Research Laboratory (ARL), as part of its mission- funded exploratory development program, has been evaluating the use of a low-frequency, ultra-wideband imaging radar to detect tactical vehicles concealed by foliage. An instrumentation-grade measurement system has been designed and implemented by ARL. Extensive testing of this radar over the preceding 18 months has led to the establishment of a significant and unique data base of radar imagery. We are currently using these data to develop target detection algorithms which can aid an operator in separating vehicles of interest from background. This paper provides early findings from the algorithm development effort. To date, our efforts have concentrated on identifying computationally simple strategies for canvassing large areas for likely target occurrences--i.e., prescreening of the imagery. Phenomenologically-sound features are being evaluated for discrimination capability. Performance assessments, in terms of receiver operating characteristics, detail detection capabilities at various false alarm rates.