Numerical Analysis of the Detection Performance

Numerical Analysis of the Detection Performance

of Ground Coupled Radars

for Different Antenna Systems and Signal Features

Davide Comite1, Alessandro Galli1, Elena Pettinelli2, and Guido Valerio3

Department of Information Engineering, Electronics and Telecommunications (DIET)

“Sapienza” University of Rome, Via Eudossiana 18, 00184 Rome, Italy

e-mail: {comite, galli}@diet.uniroma1.it 2

Department of Mathematics and Physics

“Roma Tre” University, Via della Vasca Navale 84, 00146 Rome, Italy

e-mail: pettinelli@fis.uniroma3.it

3

Institut d’Electronique et de Télécommunications de Rennes (IETR) Université de Rennes 1, Avenue du Général Leclerc, 35042 Rennes, France

e-mail: guido.valerio@univ-rennes1.fr

These specific issues are investigated in this context Abstract—The detection capabilities of ground penetrating

radar (GPR) for shallow subsurface investigations is primarily through ad-hoc analyses using a customized numerical setup related to the efficient design of the antennas and to the [5], based on a time-domain commercial CAD tool (CST suitable choice of the signal waveforms. An exploration of such

Microwave Studio) [6].

aspects is the subject of the present work, with specific interest

Thanks to the relevant advantages of the numerical to ground-coupled monostatic and bistatic configurations. In

implementation as regards economy and flexibility, a order to address in a consistent way the issues related to the

variety of simulations of GPR systems has specifically been influence of the type of antennas and of the transmitted signal

traces, extensive analyses have been performed numerically, carried out for testing different GPR antenna configurations based on ad-hoc implementation of a full-wave electromagnetic and signal features (in terms of waveforms and spectra), CAD tool. The relevant effects on the reconstruction of radar referring to typical scenarios of practical interest (such as sections are evaluated and discussed in a number of test cases

those involving GPR shallow-subsurface investigations for

for the detection of buried scatterers in scenarios of practical

planned planetary missions) [5,7-9]. interest.

To this aim, a number of different topologies of GPR

Index Terms—Ground Penetrating Radar (GPR), antenna antennas are designed and implemented, analyzing system, signal waveforms, shallow subsurface, detection of

numerically their features in terms of matching with the

buried scatterers.

external environment, bandwidth, and field distributions (both in far-field and in near-field conditions). The desirable

I. INTRODUCTION effects related to wide-band radiators are particularly

In a wide variety of geophysical, civil, and space emphasized in connection with different waveforms that can applications, the capability of properly detecting and be chosen for the GPR transmitted signal. locating ‘buried targets’ is a topic of great scientific interest. Compared quantitative analyses are thus possible for the For these purposes, different types of ground penetrating study of the scattering effects due to buried targets in radar (GPR) systems are commonly used in monostatic and various practical conditions. This allows us, for instance, to bistatic ground-coupled configurations [1,2]. test the detection performance of targets with dimensions Broadly speaking, the achievement of qualitatively valid comparable to the typical signal wavelengths and buried ‘direct’ data on the scattering effects of targets by means of very close to the interface where the ground-coupled GPR is primarily dependent on the antenna features and on antenna systems can operate. Relevant results on the the signal waveforms chosen, in addition to the scenario-scattering of targets are presented and discussed in the frame

related aspects (such as the geometry of the target and its of useful GPR applications. electromagnetic contrast with background, the possible

II. DESIGN AND TEST OF GPR ANTENNAS attenuation and dispersion in the media, the presence of

inhomogeneities, etc.). Based on the availability of the A versatile numerical implementation of a ground-direct scattering data, it is then possible to apply the most coupled system, based on CST Microwave Studio, has appropriate and efficient inversion procedures for the allowed us to efficiently design and test different types of correct detection of the buried targets [1-4]. GPR configurations.

1

The typical simulated scenario of investigation considers an inhomogeneous region basically consisting of two different media (e.g., an air/ground environment), in which arbitrary scatterers can be located and detected through monostatic or bistatic interfacial antenna systems, as sketched in Fig. 1.

0-5-10-15

S11

(dB)-20

-25

-30-35-40

0.5

0.7

0.9

1.1

1.3

1.5

(a)

S11(dB)

-10-20-30-40

0.5

1

1.5

2

2.5

3

-50

Fig. 1. Our GPR environment of interest, simulated with a time-domain electromagnetic CAD tool. The typical scenario for target detection is represented by an upper half-space (for instance, air) and a lower region describing a ground medium (represented by appropriate permittivity, permeability, and conductivity parameters). Buried scatterers having arbitrary geometry (shape, dimension, and location) and electromagnetic parameters can be considered, detectable with proper GPR antenna systems conveniently located close to the interface y = 0 (a simple Tx/Rx dipole configuration is sketched).

(b)

Fig. 2. Return loss (|s11|, dB) vs. frequency f (GHz) of various designed radiating elements for GPRs: (a) λ/2 dipole, resistively-loaded λ/2 dipole, printed folded loaded λ/2 dipole, printed monopole; in these cases a free-space external environment is considered. (b) Vivaldi antenna, radiating either in free space or coupled to a half-space dielectric ground having εr = 3.2. See labels for the colors associated to the various curves.

Among the various radiating elements that can be considered in the applications, we have designed and tested some fundamental configurations of GPR Tx/Rx systems often commonly employed in practice: e.g., i) Hertzian dipole probes; ii) resonant half-wavelength dipoles; iii) resistively-loaded λ/2 dipoles; iv) printed folded loaded λ/2 dipoles; v) printed monopoles; vi) printed Vivaldi antennas. In order to make consistent comparisons, in the following we refer to the design of antennas operating in a range centered around the reference frequency of 1 GHz. It should be emphasized that in such GPR applications it is extremely important to test the antenna features in the ‘effective’ operation conditions, that is taking into account various ‘realistic’ issues affecting the overall performance of the system, such as the nonhomogeneous external environment, the possible coupling effects between radiators, etc..

Examples concerning the matching features of the various simulated antennas are derived by calculating the return loss as a function of frequency, as shown in Fig. 2 for a number of topologies: various types of dipole and monopole configurations are in Fig. 2(a), while Vivaldi elements are in Fig. 2(b). The capability of reaching significant wideband behaviors is particularly noticeable in the latter case. In order to assess the influence of the external environment in ground-coupled configurations, in Fig. 2(b) the matching features are evaluated by considering both the radiating antenna in free space (red curve) and the antenna placed at the interface between air and a specific ground medium (blue curve).

The spatial field distribution radiated by the antennas has been evaluated considering both ‘ideal’ far-field radiation patterns and ‘real’ field configurations in typical near-field operating conditions, which take into account also the presence of the inhomogeneous environment. Fig. 3(a) shows the far field radiated in free space (Fraunhofer region) of different types of antennas (folded loaded dipole, monopole, Vivaldi radiator, compared also to an ideal current line). Fig. 3(b) shows a near-field distribution radiated at a fixed distance by different antennas (monopole, Vivaldi, and ideal current line as a reference), located at the air/ground interface (see in the relevant caption the geometrical and physical parameters).

Far Field (V/m)

E (V/m)

270

270

(a)

(b)

Fig. 3. Simulated field distributions for different GPR antenna systems, as labeled by the different colors of the curves. (a) Far-field radiation patterns (Fraunhofer region) for different antennas located in free space (polar form in a cross section xy of the scenario as in Fig. 1). (b) Near-field distribution at a fixed distance (15 cm) for different antennas located at the interface y = 0 between air (upper space) and a dielectric ground medium with εr = 3.2 (lower space).

The 8th European Conference on Antennas and Propagation (EuCAP 2014)

This basic analysis shows to what extent the antenna performance can be affected by the practical GPR operative conditions. These aspects, in conjunction with the specific features of the GPR transmitted signal, heavily influence the detection performance of the system, as discussed further.

III. WAVEFORMS AND DETECTION FEATURES The potential of our numerical implementation has been tested to obtain results for various GPR scenarios of practical interest. We refer here for instance to a ‘sandy’ dielectric ground (εr = 3.2) in which targets having different location, shape, dimensions, and contrast can be buried.

As said, among the other parameters, the choice of the signal waveform significantly influences the detection capability for scatterers in particular critical conditions [3-5]. Some test cases of typical traces considered here are: i) Gaussian-modulated pulse; ii) Ricker waveform; iii) compressed chirp signal.

Simulated results on the scattering effects of targets can be derived in the form of single gathered time-domain traces and of B-scan cross sections (‘radargrams’), both in monostatic and in bistatic configurations, analogously to the typical output achievable by GPR instruments [1,2,5].

403020

E(V/m)

100-10-20-30

2

4

6

8

10

We present in Fig. 4 some examples of test cases related to the presence of buried scatterers (see figure captions and labels for all the details). In Fig. 4(a) we show the probed output signal traces (amplitudes vs. time) from an input Gaussian pulse related to various dipole antennas.

B-scan plots (received scattered signal in grey scale vs. scan position) have been simulated for different antennas and waveforms to detect buried scatterers. Examples of detection are reported for a cubic target close to the interface in Figs. 4(b) and 4(c): in Fig. 4(b) for a monopole antenna radiating a Gaussian modulated pulse, and in Fig. 4(c) for a Vivaldi antenna radiating a chirp signal.

IV. CONCLUSION

An extensive numerical study has been carried out for the accurate analysis of the features of GPR systems in realistic scenarios. The influence of the designed antenna system and of the chosen input signal features have particularly been addressed, taking into account a number of practical issues which make this type of study an efficient and versatile tool to predict the practical performance of GPRs for target detection.

ACKNOWLEDGMENTS

The authors gratefully acknowledge financial support from the Italian Space Agency through Contract ASI I/060/10/0, EXOMARS SCIENCE phase C2/D.

REFERENCES

[1] D. J. Daniels, Ed., Ground Penetrating Radar. The Institution

(a)

t (ns)

[2] [3]

[4]

(cm)

t (ns)

(b)

[5]

(cm)

[6] [7]

(c)

[8]

Fig. 4. Simulation examples of target detection with ground-coupled radars for different antenna types and input signal waveforms. A metallic cubic scatterer, having 9-cm side and 10-cm depth, buried in a dielectric ground ( r = 3.2): (a) received traces (probed output signal E, V/m, vs. time, ns) for λ/2, loaded, and folded-loaded dipoles (see labels for the colors associated to the curves); (b) B-scan plot (grey-scale amplitude along tìme, ns, vs. linear scan position, cm) for a monopole antenna with an input Gaussian modulated pulse signal; (c) B-scan plot for a Vivaldi antenna with an input chirp signal (pulse compression ratio Bτ = 100).

[9]

of Electrical Engineers (IEE), London, UK, 2nd Ed., 2004. H. M. Jol, Ed., Ground Penetrating Radar: Theory and Applications. Elsevier, Amsterdam, The Netherlands, 2009. F. Soldovieri, I. Catapano, P. M. Barone, S. E. Lauro, E. Mattei, E. Pettinelli, G. Valerio, D. Comite, and A. Galli, “GPR estimation of the geometrical features of buried metallic targets in testing conditions,” Progress in Electromagnetic Research B, vol. 49, pp. 339-362, 2013. A. Galli, D. Comite, I. Catapano, G. Gennarelli, F. Soldovieri, and E. Pettinelli, “3D imaging of buried dielectric targets with a tomographic microwave approach applied to GPR synthetic data,” Int. J. Antennas Propag., art. ID 610389, 10 pp., 2013. G. Valerio, A. Galli, P. M. Barone, S. E. Lauro, E. Mattei, and E. Pettinelli, “GPR detectability of rocks in a Martian-like shallow subsoil: A numerical approach,” Planetary Space Sci., vol. 62, pp. 31-40, 2012.

CST Microwave Studio Manual, CST, Germany, 2002.

E. Pettinelli, G. Vannaroni, E. Mattei, A. Di Matteo, F. Paolucci, A. R. Pisani, A. Cereti, D. Del Vento, P. Burghignoli, A. Galli, A. De Santis, and F. Bella, “Electromagnetic propagation features of ground-penetrating radars for the exploration of Martian subsurface,” Near Surface Geophys., vol. 4, pp. 5-11, 2006.

E. Pettinelli, P. Burghignoli, A. R. Pisani, F. Ticconi, A. Galli, G. Vannaroni, and F. Bella, “Electromagnetic propagation of GPR signals in Martian subsurface scenarios including material losses and scattering,” IEEE Trans. Geosci. Remote Sensing, vol. 45, pp. 1271-1280, May 2007.

V. Ciarletti, C. Corbel, D. Plettemeier, P. Cais, S. M. Clifford, and S.-E. Hamran, “WISDOM GPR designed for shallow and high-resolution sounding of the Martian subsurface,” Proc. IEEE, vol. 99, pp. 824-836, May 2011.

本文来源：https://www.bwwdw.com/article/wsdm.html