Feasibility study into the identification of landmines using UWB radar: an analysis using synthesized data

The goal of this research is to detect and identify buried land mines (LM). The technique being investigated is the use of ultra-wideband (100 MHz to 3 GHz) ground penetrating radar (UWB/GPR). The first step is to gain an understanding of the scattered signal received from the use of such a system. A major point of interest is: which electromagnetic scattering features and signal processing techniques can be used in the detection/identification process? Possible candidates of scattering features are late-time resonances and diffusion poles, whereas signal processing techniques are synthetic aperture radar (SAR) imaging, correlation and/or E-pulse receivers. It is important to have good detection and identification algorithms, which work well in the ideal situation, before one can deal with the practical situation, in which there will be uncertainty in various physical parameters such as: the local electrical characteristics of the ground, the local heterogeneity of the ground (rocks, layers, tree and plant roots), surface roughness of the ground and covering foliage, position (depth and orientation) of the LM, and relative position of the radar with respect to the ground and the LM. Thus, in this work, we investigate the post-processing of impulse type radar signals, where the signals are synthesized, for relatively ideal situations, using the finite-difference time-domain (FDTD) technique.