A broadband electromagnetic induction (EMI) sensor is developed to detect and discriminate between types of metallic targets. The sensor uses a single dipole transmit coil and two dipole receive coils in a differential configuration. The sensor operates in the frequency domain and collects data at eight logarithmically spaced frequencies from 105 Hz to 1.905 kHz. The system is configured for relatively distant targets in fairly conductive soil or water. The system is designed to be relatively lightweight while being sensitive and making accurate measurements. Experimental results are presented for a prototype sensor.

Many different coil head configurations are used in electromagnetic induction (EMI) systems for sensing buried targets, but these coils are most often designed by altering known winding configurations because a general wire parameterization is difficult to create and to optimize for target sensitivity. The soil sensitivity, which is important in mineralized soils, is also often not taken into account. This work presents a method of parameterizing wire coils using stream functions. A new set of normalized metrics for analyzing the coils as stream functions is demonstrated, and then the metrics and stream functions are optimized using a new biconvex optimization procedure. The stream functions are then converted back into wire coils to demonstrate the efficacy of the stream-function parameterization. Results show an improvement in target and soil sensitivity performance over known conventional wire coil configurations.

We present three analyses that explore the behavior of the Cramer–Rao lower bounds (CRBs) on estimating a target’s relaxation frequency response from frequency-domain electromagnetic induction (EMI) data. In the first analysis, we show that the CRB on a given relaxation frequency is independent of the amplitudes of the other relaxations present in the target’s relaxation frequency response. In the second, we show that the presence of a second relaxation frequency closer than one decade in frequency greatly increases the CRB on that relaxation frequency. Finally, we illustrate the behavior of the minimum root-mean-square error (RMSE) per relaxation frequency as a function of the number of relaxation frequencies present in the target’s relaxation frequency response.

We derive the Cramer–Rao lower bounds (CRBs) for all target parameters associated with noisy measurements of a specific class of targets using a frequency-domain electromagnetic induction (EMI) system. The target parameters include the target tensors, the relaxation frequencies and their corresponding amplitudes, as well as the target location. We validate the derivation through Monte Carlo simulation. We then derive approximate CRB (ACRB) expressions based on a new low-rank model perspective for EMI data. These approximate expressions are significant simplifications over the full CRB expressions. They also facilitate the analysis and improve the understanding of the factors impacting the lower bounds on the target parameters. We demonstrate the utility and accuracy of the ACRB expressions for two example targets.

Sinuous antennas are capable of producing ultra-wideband radiation with polarization diversity in a low-profile form factor, making them a good fit for close-in sensing applications such as ground-penetrating radar (GPR). This work proposes an unconventional method of operating a four-port sinuous antenna-driving each arm independently and unbalanced-to achieve a quasi-monostatic antenna system capable of polarimetry while separating transmit and receive channels, as is common in GPR systems. The quasi-monostatic configuration of the antenna reduces system size as well as increasing sensitivity to near-surface targets by preventing extreme bistatic angles. A prototype four-port sinuous antenna is fabricated and integrated into a GPR testbed. The polarimetric data obtained with the antenna is then used to distinguish between buried target symmetries.

An electromagnetic induction (EMI) sensor is investigated that uses a spinning magnet to generate the primary magnetic field. The major advantage of the technique is power efficiency as no power is needed to establish the field, and very little power is needed to keep the magnet spinning. The sensor uses two receive coils in a differential configuration to make it much less sensitive to ambient noise relative to a prior sensor that uses a single receive coil. Even with the differential coils, the sensor has a stronger response than the prior single-coil sensor for target ranges of most interest. The sensor has an interesting phase response that will help in refining the location of a target. Experimental results are presented for a prototype sensor.

Electromagnetic induction (EMI) sensors are widely used to detect magnetic and/or conductive objects that are concealed in some manner. The sensors generate a time-varying magnetic field that is used to excite a dipole moment on the object which generates a secondary magnetic field that is used to detect the object. A coil driven by a time-varying current is used almost exclusively to generate this magnetic excitation. For practical coils, the power dissipated due to resistive losses can be problematic especially for small coils. The use of a spinning magnet to make the magnetic excitation is investigated as a method for reducing the power dissipation. Large potential power savings are demonstrated using commercial neodymium magnets for a relatively low-frequency EMI system.

Wideband electromagnetic induction sensors have been shown to be more effective than traditional metal detectors at identifying metallic targets of interest when buried near abundant metallic clutter. A promising strategy for performing target identification using wideband data relies on characterizing the target response using a magnetic polarizability tensor (MPT) that is a function of frequency. In this paper, a surface integral method is presented for deriving the MPT in singularity expansion form using a modal decomposition. The method is verified by comparing the singularity expansion for a spherical shell to a derived analytical solution. The expansion coefficients for cylindrical tubes, of various aspect ratios, are compared to experimental results, showing good agreement. Finally, expansion coefficients are given for a disk and are compared to other numerical results.

We derive the Cramer-Rao lower bound (CRB) for target tensor amplitudes and target location parameters that can be estimated from electromagnetic induction (EMI) measurements of a target. In deriving the bound, no restrictions are placed on the target type, target orientation, quantity, or location of the measurement positions, nor the geometry of the EMI sensor or sensor array. The analysis is applicable to both a scanned sensor and a stationary array, as well as both time- and frequency-domain sensors. We show how the bound varies as a function of target type, orientation, depth, and signal-to-noise ratio. In addition, we illustrate the ways to use the CRB and the Fisher information matrix to analyze the relationships between the tensor amplitudes and location parameters. We show results of the CRB analysis applied to an experimental frequency-domain Georgia Tech sensor for a few target types for a nominal 2-D scan geometry as well as the Time-domain Electromagnetic Multi-sensor Towed Array Detection System. We also provide an algorithm for estimating the target location and tensor that achieves the CRB.

Broadband electromagnetic induction sensors have been shown to be highly effective at detecting and classifying buried metal. These sensors generate a wideband primary magnetic field to excite eddy currents in buried metal and measure the secondary magnetic field they induce. The target is identified by fitting a dipole model to the measured data. The dipole model is represented by a magnetic polarizability tensor, which is expressed by a pole expansion and used as a signature for a target. Viewed as a singularity expansion, the magnetic polarizability can be decomposed into contributions from a set of eddy-current modes that relax at different frequencies. This decomposition provides a unique signature for targets of interest, allowing them to be effectively distinguished from clutter. In this paper, an integral method is presented for computing the magnetic polarizability and its singularity expansion for three-dimensional targets.

An error mechanism for continuous wave electromagnetic induction sensors has been discovered and analyzed. The error is due to the signal processing when the received signal is modulated by the relative motion between the target and the sensor. A simple modification of processing the data has been developed that mitigates this error, and the method is demonstrated with experimental data.

This work presents a mathematical approach for recovering a physical model from a low-rank approximation of measured data obtained via the singular value decomposition (SVD). The general form of a low-rank physical model of the data is often known, so the presented approach learns the proper rotation and scaling matrices from the singular vectors and singular values of the SVD in order to recover the low-rank physical model of the data from the SVD approximation. By recovering the low-rank physical model, it becomes possible to exploit specific knowledge of the model to extract meaningful information for the physical application being studied. This work is useful for processing wide-band electromagnetic induction data-the motivating application.

Software defined radio (SDR) is a rapidly developing technology that implements signal processing components partially or completely in software. In this paper, SDR's potential as a platform for ground-penetrating radar (GPR) is explored. The stepped-frequency radar method is implemented using off-the-shelf SDR hardware and open-source software. SDR is typically designed for communications applications, so special consideration is necessary for remote sensing. Precisely timed commands achieve RF phase coherence as well as digital sample synchronization. A multi-tone digital signal takes advantage of instantaneous bandwidth, and a sequential tuning routine expands effective bandwidth to cover the 500-5000 MHz band. Initial design challenges are weighed against potential advantages of design flexibility and hardware versatility.

Broadband electromagnetic induction sensors can effectively detect and classify buried metal, such as landmines, by using an incident magnetic field to induce eddy currents in buried metal and measuring the secondary magnetic field they produce. A target's magnetic polarizability is a frequency dependent tensor that expresses the relationship between the incident magnetic field and the scattered field created by the eddy currents. Viewed as a singularity expansion, the magnetic polarizability can be decomposed into contributions from a set of eddy-current modes which relax at different frequencies. This decomposition provides a unique signature for targets of interest, allowing them to be effectively distinguished from clutter. In this paper, an improved, stream-function-based numerical approach will be presented that can accurately compute the eddy-current modes that flow in thin conducting shells and their contribution to the singularity expansion of the magnetic polarizability.

Comparing the performance of physical coil heads used in EMI systems is generally straightforward. Dimensions, wire diameters, transmit power, amplifier noise, winding patterns, etc., are known, so target and soil sensitivity can be easily measured or calculated and then compared. However, comparing the relative performance of different coil head configurations - i.e. only winding patterns without influence from other factors - that are used for electromagnetic induction sensing is a non-trivial task. Metrics for both coil head sensitivity and soil sensitivity that allow comparison of coils based only on their winding patterns are developed. As an example of their use, these metrics are then used to choose the best double-D coil head from a range of configurations.

Wideband electromagnetic induction (WEMI) sensors are used to detect, classify, and locate obscured targets. WEMI sensors can be designed to measure the magnetic fields either in the time domain or frequency domain. They also capture multiple location measurements from a target by either a physical sensor array, a synthetic sensor array created by scanning the sensor, or a combination of the two. This letter uses a physical model for the sensor to develop a low-rank model for WEMI data. The data are organized into a matrix where the rows contain the time/frequency measurements and the columns contain the multiple locations so all of the data can be exploited jointly. It is shown that this data matrix can be processed directly with singular value decomposition (SVD) to extract three independent terms-one for target signature, one related to location, and one for tensors that define the orientation. The low-rank model predicts a maximum rank for WEMI measurements and is exploited to provide a relationship between the rank of the measurements and the number of linearly independent tensors for the target. Results are presented from the laboratory data to validate the low-rank model and demonstrate the connection between the data's rank and the physical target. The low-rank model leads to a new “filterless” processing paradigm for exploiting WEMI data by using the SVD to perform sensor calibration, hardware debugging, as well as target detection.

Electromagnetic induction sensors excel at detecting and classifying buried metallic objects. With the help of precomputed electromagnetic models, detection algorithms are able to pinpoint a target's location and orientation with high accuracy. A target's discrete spectrum of relaxation frequencies, is particularly useful for detection purposes. The discrete spectrum of relaxation frequencies is a simple pole expansion of the target's reaction to magnetic excitations. Its pole expansion coefficients can be computed by formulating differential Maxwell's equations as a generalized eigenvalue problem. When modeling arbitrary three-dimensional objects, this approach is complicated by the size of the system matrices and by a large null space that is near the eigenvalues of interest. A modified Jacobi-Davidson iteration will be described in this paper, that is capable of extracting the eigenvalues of interest without interference from the null space.

Continuous-wave (CW) electromagnetic induction (EMI) systems operating in the presence of magnetic soil often encounter issues with the voltage that the soil induces in the receive coil. Previously, an optimization procedure that represents the coils as stream functions and attempts to create coils to mitigate the effects of the soil was presented. In this paper, the optimization convergence is improved, and a new soil constraint that improves the coils created by the optimization is developed. New coil metrics are introduced, and new coils are designed using the new soil constraint, which allows more freedom for the optimization to adjust the soil response. The new coils indicate the possibility of improvement over current coil designs.

A method using feedback is presented that reduces several measurement errors inherent in electromagnetic induction sensors. Errors associated with coupling between receive coils and errors associated with operating near magnetic soils will both be reduced. The method uses feedback that is directly injected into the receive coils and does not require secondary coils. A simple circuit is introduced to perform the feedback and is optimized to reduce the errors and make the circuit stable. Experimental results are presented to show the effectiveness of the feedback.

Low-frequency, broadband electromagnetic induction (EMI) sensors have been shown to be highly effective at detecting and classifying buried conducting targets. Even the simplest inductive sensors are capable of easily detecting the presence of buried metal. It is difficult, however, for most inductive sensors to discriminate between targets of interest, and the ubiquitous metallic clutter that might be buried alongside it. One attractive solution to this discrimination problem is to invert the broadband frequency data collected by the EMI sensor, finding a pole-expansion representation of the scattering transfer function known as its discrete spectrum of relaxation frequencies (DSRF). A target's low-frequency scattering behavior can then be compactly described by a set of relaxation frequencies that are the poles of the transfer function and their coefficients that are the corresponding amplitudes. Discrimination can be achieved by comparing the inverted data to a dictionary that contains the theoretical DSRFs of targets of interest. In this paper, a computational method will be presented for modeling the DSRF of arbitrarily-shaped three-dimensional conducting targets.

Continuous-wave (CW) electromagnetic induction (EMI) systems operating in the presence of magnetic soil often encounter issues with the voltage that the soil induces in the receive coil. A formulation is developed to allow the calculation of the soil response of a CW EMI coil head that is represented by stream functions. This formulation is then included in an optimization procedure for stream-function coils that is designed to optimize a coil head for improved sensitivity while nulling the coil head's soil response at a specific height above the soil. Example coils are created using the optimization method, and the new coils are compared to other common coil types.

A method using magnetic feedback is presented that reduces measurement errors in an electromagnetic induction sensor operating near magnetic soils. The method uses a feedback coil that is placed near the receive coil in the sensor to partially cancel the errors introduced by the magnetic properties of the soil. A soil sensitivity metric is introduced to quantify the effects of the soil, and this metric is used to optimize a circuit used to drive the feedback coil. Theoretical and experimental results are presented to show the effectiveness of the feedback.

Recent developments in the estimation of the discrete spectrum of relaxation frequencies (DSRFs) has opened doors to more robust subsurface target discrimination using electromagnetic induction measurements. In particular, a nonnegative least squares DSRF (NNLSQ-DSRF) estimation method has been shown to be robust and free from parameter tuning. In this letter, we propose an adaptive prefiltering process to complement the NNLSQ-DSRF where we attempt to linearly combine measurements and produce a filtered signal that is very likely to have a nonnegative DSRF, as well as an enhanced signal-to-noise ratio. Using synthetic and field data, we demonstrate that the proposed adaptive prefilter can effectively produce signals with nonnegative DSRFs.

Ground-penetrating radar (GPR) is used to image and detect subterranean objects, for example, in landmine detection. Although full 3-D inversion of GPR measurements is possible for simple algorithms such as backprojection, it is impractical when using more advanced algorithms that involve ℓ1- minimization. Many of the algorithms used for GPR imaging involve the storage, or online generation, of a huge dictionary matrix created from discretizing a high-dimensional nonlinear model. This parametric model includes all the target features that need to be extracted, including 3-D location, object orientation, and target type. As more parameters are added to the model, the dimensionality increases. If uniform sampling is done over high-dimensional parameter space, the size of the dictionary and the complexity of the inversion algorithms rapidly grow, exceeding the capability of real-time processors. This paper shows that strategic structuring of the dictionary, which takes advantage of translational invariance in the model, can reduce the dictionary storage by several orders of magnitude and exploit the fast Fourier transform for fast computation of previously highly impractical, bordering on impossible, 3-D GPR imaging problems.

The ability to detect and dispose of buried mines requires effective means by which to discriminate between hazardous targets and benign clutter. In that regard, wide-band electromagnetic induction (EMI) sensors have shown significant promise in their ability to classify buried metallic objects based on their response to illumination by a time-varying magnetic field. A target's scattered response may be expressed compactly in its magnetic polarizability dyadic, a form that describes the reaction of the scatterer to an arbitrary magnetic field. The magnetic polarizability dyadic may be written in terms of the eddy currents that are induced in the target. The method described in this paper uses a scalar stream function as a basis for the eddy currents that are induced in the target. This approach is powerful since the solenoidality of the current density is enforced trivially and its boundary conditions may be enforced elegantly. By setting up the eddy current equation as a generalized eigenvalue problem we arrive at a modal decomposition of the polarizability dyadic. Distribution A: Approved for public release.

Electromagnetic induction (EMI) sensors are commonly used to detect and locate buried metallic objects such as landmines, but they are capable of extracting much more information about objects, e.g., location and magnetic polarizability. Recent research has led to an effective inversion method to extract target orientation by finding the tensor representation of the target [1-3]. The “tensor amplitude” extraction techniques also have an important capability that has not been fully examined until now, and that is the ease with which they can be used with a variety of EMI sensor geometries. This paper will examine how making slight alterations to an existing sensor geometry can dramatically increase its effectiveness, while still using the “tensor amplitude” extraction to accurately and efficiently determine the unknown parameters of a metallic object.

Continuous-wave electromagnetic induction systems used for subsurface sensing often employ separate transmit and receive coils. In these systems, it is desirable to have zero mutual coupling between the transmit and receive coils and for the coils to have maximum sensitivity at a specific location. A representation of a pair of coils as stream functions on planar surfaces is created, and then a method of optimizing these stream functions for minimum mutual coupling and maximum sensitivity using a convex solver is developed. A pair of coils is optimized for varying amounts of dissipated power and stored energy, and the solutions are categorized. The sensitivity of the optimized coils is then compared to other common coil types.

Sensor array measurements can be inverted to image a region containing targets. The resulting amplitude image is usually interpreted as target strength versus location, but often the imaged amplitude is a function of more parameters than just the location. Sparse target regions can be imaged with dictionary based modeling which relies on enumeration of each parameter with a dense grid. With many parameters, the dictionary becomes too large, which leads to computational complexity issues. This paper shows how additional parameters, such as target orientation and symmetry, can be represented by a tensor matrix instead of a simple amplitude. Furthermore, the tensor can be treated as a continuous variable just like amplitude, which enables extraction of multiple parameters, while reducing the storage requirements of the dictionary, and reducing off-grid modeling error.

Electromagnetic band gap (EBG) antennas may be suitable for handheld sensing applications, like ground penetrating radar (GPR) for landmine detection, because of their small size, efficiency, and directivity. Increased detection performance has been shown when a GPR is combined with a metal detector, but a typical EBG antenna would cause a response in the metal detector because of the large amount of metal in the structure. An EBG composed of very thin metal is proposed in this paper for application as a GPR co-located with a metal detector without causing a significant response in the metal detector. Manufacturing methods are discussed and GPR measurements are shown using the thin-metal EBG antennas. The metal detector response from a thin metal sheet is discussed and measurements of the EBG ground planes are shown using a laboratory wideband electromagnetic induction system.

Electromagnetic induction (EMI) systems often use separate transmit and receive coils. In these systems, it is desirable for the transmit and receive coils to have both minimal mutual coupling and a maximum field product, thus maximizing the detection depth. A mathematical representation is chosen for a pair of spiral coils that allows the coils to be optimized using an iterative convex method. This method, which is very fast, allows the coils to be optimized for the desired properties of minimum mutual coupling and a maximum field product. We then present results showing a pair of nonuniformly-wound, double-sided spiral coils.

Broadband, continuous-wave, electromagnetic induction sensors have been shown to perform well, but they usually use more power than a pulsed induction sensor because of the power wasted in the linear amplifier used to drive the coils. Generally, much more power is dissipated in the power amplifier than in the coils in these systems. Methods for reducing this power are investigated using both linear and switched-mode amplifiers. Significant reductions in the power dissipated by the linear amplifier are achieved by optimizing the signal used to drive the coil. Much greater reductions are achieved with a switched-mode amplifier using an optimized signal that has most of its energy in the desired frequencies for the sensor. Both of these techniques are shown to perform well in a prototype system.

Electromagnetic band-gap (EBG) antennas are ideal for handheld sensing applications, such as ground-penetrating radar (GPR) for landmine detection, because of their small size, efficiency, and directivity. Increased detection performance has been shown when a GPR is combined with a metal detector, but a typical EBG antenna would preclude the sensors from being colocated because of the large amount of metal in the EBG structure. An EBG composed of very thin metal is proposed in this letter for application as a GPR colocated with a metal detector, without causing a significant self-response in the metal detector. Manufacturing methods are discussed, and GPR measurements are shown from the thin-metal EBG antennas. The metal detector response from a thin metal sheet is discussed, and measurements of the EBG ground planes are shown using a laboratory wideband electromagnetic induction system. The response from the thin metal is shown to be as low as three orders of magnitude less than from a copper sheet at typical metal detector frequencies.

Frequency-domain electromagnetic induction (EMI) sensors can measure object-specific signatures that can be used to discriminate landmines from harmless clutter. In a model-based signal processing paradigm, the object signatures can often be decomposed into a weighted sum of parameterized basis functions, such as the discrete spectrum of relaxation frequencies (DSRF), where the basis functions are intrinsic to the object under consideration and the associated weights are a function of the target-sensor orientation. The basis function parameters can then be used as features for classifying the target. One of the challenges associated with effectively utilizing a model-based signal processing paradigm such as this is determining the correct model order for the measured data, as the number of basis functions containing fundamental information regarding the target under consideration is not known a priori. In this paper, sparse Bayesian relevance vector machine (RVM) regression is applied to simultaneously determine both the number of parameterized basis functions and their relative contributions to the measured signal assuming a DSRF signal model. The target is then classified utilizing the basis function parameters as features within a statistical classifier. Results for data measured with a prototype frequency-domain EMI sensor at a standardized test site are presented, and indicate that RVM regression followed by distance-based statistical classifiers utilizing the resulting model-based features provides an effective approach for classifying and identifying landmine targets.

An iterative reweighted algorithm is proposed for the recovery of jointly sparse vectors from multiple-measurement vectors (MMV). The proposed MMV algorithm is an extension of the iterative reweighted ℓ1algorithm for single measurement problems. The proposed algorithm (M-IRL1) is demonstrated to outperform non-reweighted MMV algorithms under noiseless measurements. A regularization of the M-IRL1 algorithm is also proposed to accommodate noise. The ability to robustly handle noise is demonstrated through an electromagnetic induction application.

A conducting body's Discrete Spectrum of Relaxation Frequencies (DSRF) is a valuable classifying characteristic that is used in the detection and discrimination of buried objects. In physical terms, the DSRF depicts the decay of eddy currents in a conducting body that is exposed to a time-varying magnetic field. An object's DSRF can only be computed analytically for a few canonical cases. Numerical computations for bodies of revolution (BOR) have been presented using boundary-element and finite-element techniques, but these techniques are complex. In this work, a simpler approach based on eigenvalue decomposition is presented that is capable of computing the DSRF for a class of BORs. Approved for public release, distribution unlimited.

In this paper, a numerical method for calculating the wideband, low-frequency, eddy-current response of a metallic body-of-revolution is described. Solutions are given in a modal decomposition that can be related to the discrete spectrum of relaxation frequencies and the measured magnetic polarizability dyadic. Calculated responses are compared with the analytically known solution for a sphere and measured responses for a cylinder. A graph of the relaxation frequencies and moments for the two most dominant modes in a cylinder, given as a function of aspect ratio, is provided.

Broadband electromagnetic induction (EMI) sensors have been shown to be able to reduce false alarm rates and increase the probability of detecting landmines. To aid in the development of these sensors and associated detection algorithms, a testing facility and inversion technique have been developed to measure the magnetic response of soils and the magnetic polarizability of metal targets. When theoretical predictions are possible, these measurements show good agreement with theory.

The EMI response of a target can be accurately modeled by a sum of relaxations. However, it is difficult to obtain the model parameters from measurements when the number of relaxations is unknown. We have previously proposed estimation methods for the model parameters from single measurements. In this paper, we exploit the invariance property of the relaxation frequencies and propose to obtain more accurate estimates using multiple measurements that are often available. This is accomplished by casting the modeling problem into a jointly-sparse vector recovery problem. The proposed method is shown to deliver robust estimation using synthetic, laboratory data, and field data.

A technique for calibrating broadband electromagnetic induction (EMI) sensors is presented. The technique is very simple and uses a powdered ferrite core as a calibration standard. The purpose of the calibration is to improve the accuracy of the senor which enhances its ability to discriminate between different types of targets.

Several landmine detection techniques using electromagnetic induction (EMI) sensors have been proposed in the past decade. In this paper, we propose a class of detection techniques based on the discrete spectrum of relaxation frequencies (DSRF). Two DSRF detection methods are demonstrated: one using the support vector machine and one using the k-nearest neighbor method. A soil model is also proposed to identify EMI response from the magnetic properties of the soil. A detection framework is suggested to incorporate the soil model and the classifier. The robustness of landmine detection using the DSRF is demonstrated.

The electromagnetic induction response of a target can be accurately modeled by a sum of real exponentials. However, in practice, it is difficult to obtain the model parameters from measurements. We previously proposed a constrained linear method that can robustly estimate the model parameters when they are nonnegative. In this letter, we present a modified$\ell_{p}$-regularized least squares algorithm, for$0 \leq p \leq 1$, that eliminates the nonnegative constraint. An empirical method for choosing the regularization parameter is also studied. Using tests on synthetic data and laboratory measurements, the proposed method is shown to provide robust estimates of the model parameters in practice.

A framework is presented for information-theoretic sensor management for the detection of static targets. The sensor manager searches for targets within a cell grid using a suite of sensor platforms. Each sensor platform may contain one or more sensing modalities, and each of these modalities has known probabilities of detection and false alarm and also has an associated cost of use. Additional information such as motion constraints on the sensors and the prior distribution of the targets in space is incorporated. The sensor manager then directs the movement of the sensors through the grid by maximizing the expected information gain that will be obtained with each new sensor observation. Key modeling questions are addressed, including the selection of an appropriate information measure and the joint or independent management of the sensors. Through a number of simulations, the performance of the sensor manager is compared to the performance of a blind sweep procedure, a random search procedure, and an alternative information-theoretic sensor manager. The intelligent sensor management procedure is demonstrated to achieve a superior performance compared to all of the other three techniques. A specific application area for which the sensor management problem is becoming more critical is landmine detection; thus, the performance of the sensor manager is also analyzed using real data from three different landmine detection sensing modalities, and the proposed sensor management technique is again demonstrated to be superior compared to more simplistic approaches.

It is difficult to robustly estimate the parameters of an additive exponential model from a small number of frequency-domain measurements, especially when the model order is unknown and the parameters must be constrained to be real. Recent work in sparse sampling and sparse reconstruction casts this problem as a linear dictionary selection problem by densely sampling the parameter space. We present a modified ℓp-regularized least squares algorithm, for 0 ≤ p ≤ 1, and show that it is effective when the frequency sampling is sparse over a couple of decades and the parameters must be estimated over more than four decades. An empirical method for choosing the regularization parameter is also studied. Using tests on synthetic data and laboratory measurements for an EMI application, the proposed method is shown to provide robust estimates of the model parameters up to eighth order.

In this paper, a technique is developed to model ground penetrating radar interactions. It combines reciprocity with the results from a full numerical model and allows one to compute a large number of scattering responses simultaneously, provided the scatterers are small. The results of an example calculation are shown and compared with a full numerical model and a possible application is demonstrated.

Broadband electromagnetic induction (EMI) sensors have been shown to be able to reduce false alarm rates and increase the probability of detecting landmines. To aid in the development of these sensors and associated detection algorithms, a testing facility and inversion technique have been developed to characterize the response of typical targets and clutter objects as a function of orientation and frequency.

The electromagnetic induction response of a target can be accurately modeled by a sum of real exponentials. However, it is difficult to obtain the model parameters from measurements when the number of exponentials in the sum is unknown or the terms are strongly correlated. Traditionally, the time constants and residues are estimated by nonlinear iterative search. In this paper, a constrained linear method of estimating the parameters is formulated by enumerating the relaxation parameter space and imposing a nonnegative constraint on the parameters. The resulting algorithm does not depend on a good initial guess to converge to a solution. By using tests on synthetic data and laboratory measurement of known targets, the proposed method is shown to provide accurate and stable estimates of the model parameters.

To better study the ground penetrating radar (GPR) problem, an important step is to develop an accurate simulation of the fields induced by the radiating antennas. In this work, a finite-difference time-domain (FDTD) model of a spiral-antenna GPR system was developed and verified against measurements of a prototype system. The paper discusses difficulties in modeling the antenna elements and presents evidence that the model properly predicts the response of the antennas to scatterers in the presence of the ground.

The EMI response of a target can be accurately modeled by a sum of real exponentials. However, it is difficult to obtain the model parameters from measurements when the number of exponentials in the sum is unknown. In this paper, a constrained linear method for estimating the parameters is formulated by enumerating the relaxation parameter space and imposing a nonnegative constraint on the parameters. Using tests on synthetic data and laboratory measurement of known targets the proposed method is shown to provide accurate and stable estimates of the model parameters. The estimated parameters are then used to cluster different targets types for classification.

Compressive Sensing (CS) uses a relatively small number of non-traditional samples in the form of randomized projections to reconstruct sparse or compressible signals. The Hough transform is often used to find lines and other parameterized shapes in images. This paper shows how CS can be used to find parameterized shapes in images, by exploiting sparseness in the Hough transform domain. The utility of the CS-based method is demonstrated for finding lines and circles in noisy images, and then examples of processing GPR and seismic data for tunnel detection are presented.

A new data acquisition and imaging method exploiting the sparsity of the target space is presented for ground penetrating radar (GPR) imaging. Sparsity is enforced by solving a convex l1minimization problem which uses a very small number of random measurements. The method can greatly reduce the data acquisition time while producing sparse target space images. Simulation and experimental data results are provided to show that the method has excellent resolution and is robust to noise and random spatial sampling.

A broadband electromagnetic induction (EMI) sensor is developed to help discriminate between buried landmines and metal clutter. The detector uses a single dipole transmit coil and an array of three quadrapole receive coils. The sensor operates in the frequency domain and collects data at 21 logarithmically spaced frequencies from 300 Hz to 90 kHz. Experimental results are presented for several targets.

The Radon Transform (RT) is known to be effective in detecting lines in noisy images, but it is not capable of detecting curves unless the curve parametrization is given. In this paper, partial Radon transforms (PRT) are investigated as a tool to detect curved features such as underground tunnels in ground penetrating radar (GPR) images. The algorithm applies the Radon Transform to small batches of the total image and updates the tunnel position parameters as new batches are used. Missing data, as well as finding the ends of tunnels can be handled with the proposed algorithm. Performance analysis is given for various signal-to-noise ratios (SNR) and batch sizes. The effect of the curvature level on the performance is also analyzed.

A broadband electromagnetic induction (EMI) sensor is developed to help discriminate between buried land mines and metal clutter. The detector uses simple dipole transmit and receive coils along with a secondary bucking transformer to mostly cancel the coupling between the coils. The technique allows the cancellation that can be obtained using a quadrupole receive coil while maintaining the depth sensitivity and simple detection zone of a dipole coil. Experimental results are presented for several targets.

We consider the problem of detecting and locating subsurface objects by using a maneuvering array that receives scattered seismic surface waves. We demonstrate an adaptive system that moves an array of receivers according to an optimal positioning algorithm that is based on the theory of optimal experiments. The goal is to minimize the number of distinct measurements (array movements) needed to localize objects such as buried landmines. The adaptive localization algorithm has been tested using data collected in a laboratory facility. The performance of the algorithm is exhibited for cases with one or two targets and in the presence of common types of clutter such as rocks in the soil. Results are also shown for a case where the propagation properties of the medium vary spatially. In these tests, the landmines were located using three or four array movements. It is envisioned that future systems could incorporate this new method into a portable mobile mine-location system

A multistatic ground-penetrating radar (GPR) system has been developed and used to measure the response of a number of targets to produce data for the investigation of multistatic inversion algorithms. The system consists of a linear array of resistive-vee antennas, microwave switches, a vector network analyzer, and a 3-D positioner, all under computer control. The array has two transmitters and four receivers which provide eight bistatic spacings from 12 to 96 cm in 12-cm increments. Buried targets are scanned with and without surface clutter, which is a layer of rocks whose spacing is empirically chosen to maximize the clutter effect. The measured responses are calibrated so that the direct coupling in the system is removed, and the signal reference point is located at the antenna drive point. Images are formed using a frequency-domain beamforming algorithm that compensates for the phase response of the antennas. Images of targets in air validate the system calibration and the imaging algorithm. Bistatic and multistatic images for the buried targets are very good, and they show the effectiveness of the system and processing.

The problem of adaptive multimodality sensing of landmines is considered based on electromagnetic induction (EMI) and ground-penetrating radar (GPR) sensors. Two formulations are considered based on a partially observable Markov decision process (POMDP) framework. In the first formulation, it is assumed that sufficient training data are available, and a POMDP model is designed based on physics-based features, with model selection performed via a variational Bayes analysis of several possible models. In the second approach, the training data are assumed absent or insufficient, and a lifelong-learning approach is considered, in which exploration and exploitation are integrated. We provide a detailed description of both formulations, with example results presented using measured EMI and GPR data, for buried mines and clutter

To investigate the problem of detecting and imaging underground tunnels, an experimental system that utilizes seismic waves has been constructed. Seismic reflections from the tunnel are transformed into a 3D image using a synthetic aperture time-delay backprojection algorithm. Results from experimental data show that the tunnel is directly visible in the backprojected image. Nevertheless, tunnels with low signal to noise ratio (SNR) are located using 2D and 3D Radon Transforms followed by a detection algorithm. A simulation is performed on the performance of the Radon transform for detecting lines in noisy images and it is shown how lines in very low SNR images can be detected. Also it is observed that longer lines have higher probability of detection at the same noise level.

An experimental system to collect co-located ground penetrating radar (GPR) and seismic data was developed to investigate possibilities of using the sensors individually or in a cooperative manner to detect shallow tunnels. These sensors were chosen because they sense very different physical properties. The seismic sensor is sensitive to the differences between the mechanical properties of a tunnel and the soil while the GPR is sensitive to the dielectric properties. Raw and processed data from both sensors are presented.

The imaging of subsurface targets using ground penetrating radar (GPR) is becoming an increasingly important area of research. Conventional image formation techniques expend large amounts of computation to resolve a region fully, even a region of clutter. However, by using multi-resolution techniques, e.g, quadtree algorithms, potential targets and clutter can be discriminated in a computationally efficient way. Prior work has focused on the development of 2D quadtree algorithms for surface targets. For mine detection, target depth adds another dimension; thus, we have developed a 3D quadtree algorithm, and applied a multi-stage detector that uses the energy change between quadtree stages to discriminate target and clutter regions. This algorithm is then tested on computer-generated data, as well as experimental data collected from a model mine field. Results show that target location information can be obtained even under near field and small aperture conditions.

An experimental system to collect co-located ground penetrating radar (GPR), electromagnetic induction (EMI), and seismic data was developed to investigate the possibility of using the sensors in a cooperative manner and to investigate the benefits of the fusion of the sensors. These sensors were chosen because they can sense a wide range of physical properties. The seismic sensor is sensitive to the differences between the mechanical properties of a landmine and the soil while the GPR is sensitive to the dielectric properties, and the EMI sensor is sensitive to the conductivity.

In this paper, the stability behavior of the first-order finite-difference time-domain algorithm for elastodynamics at the interface between two different materials is investigated. A necessary condition for stability is established, which, dependent on the material properties of the two media, might be more restrictive than the well-known Courant condition. It is shown that this more restrictive stability condition can be avoided if the material properties are averaged on the boundary.

A three-dimensional (3-D) finite-difference model for elastic waves in the ground has been developed and implemented. The model has been created to supplement the development of a sensor that uses elastic waves to detect buried land mines. The model is used to investigate the propagation characteristics of elastic waves in the ground and to explore the interaction of elastic waves with buried land mines. When elastic waves interact with a buried mine, a strong resonance occurs at the mine location. The resonance can be used to enhance the mine's signature and to distinguish the mine from clutter. Results presented in this paper explain the features of elastic wave propagation in the ground and show the interaction of elastic waves with both an anti-personnel mine and an anti-tank mine.

A laboratory-scale experimental model has been developed and tested for a system that uses artificially generated high-frequency seismic waves in conjunction with a radar-based noncontact displacement sensor to detect buried landmines. The principle of operation of the system is to measure the transient displacement field very close to a mine location. In this way, the absorption and the geometrical spreading of the seismic waves have not reduced the effects of the mine. By using a seismic excitation, the system exploits the large difference between the elastic properties of a mine and the surrounding soil. This difference causes seismic wave interactions in the vicinity of a mine to be quite distinctive and provides a method for imaging mines and distinguishing them from typical buried clutter. Images of a variety of simulated and inert anti-tank and anti-personnel mines have been formed using this system. Burial scenarios involving natural clutter (rocks and sticks), light surface vegetation, localized burial effects, and multiple mines in close proximity have been studied. None of these scenarios appears to pose serious problems for detection performance.

A two-dimensional (2-D) finite-difference model for elastic waves in the ground has been developed. The model uses the equation of motion and the stress-strain relation, from which a first-order stress-velocity formulation is obtained. The resulting system of equations is discretized using centered finite-differences. A perfectly matched layer surrounds the discretized solution space and absorbs the outward traveling waves. The numerical model is validated by comparison to an analytical solution. The numerical model is used to study the interaction of elastic waves with a buried land mine. It is seen that the presence of an air-chamber within the mine gives rise to resonant oscillations that are clearly visible on the surface above the mine. The resonance is shown to be due to flexural waves being trapped within the thin layer between the surface of the ground and the air chamber of the mine. The numerical results are in good qualitative agreement with experimental observations.

The measured electrical constitutive parameters (effective relative permittivity and effective conductivity) for Georgia red clay are presented. These results are for the frequency range 50 MHz-1.25 GHz and for six samples with water contents by dry weight ranging from approximately 0-30%. These electrical parameters should be useful in designing and characterizing broadband systems whose performance is dependent upon or affected by the Earth, such as ground-penetrating radars.>