Multiple moving target detection with ultra wideband radar using super-resolution algorithms

Official URL: http://risc01.sabanciuniv.edu/record=b1169203 (Table of Contents)

The improvements in microwave electronics opened the way to build microwave components such as low noise amplifiers, samplers and pulse generators that are broadband. As these building blocks are being developed, new applications become subject of research. Ultra wideband radar is one of these subjects. Major applications of ultra wideband radars are behind the wall imaging, biomedical imaging and buried land mine detection. In this study we aimed to locate multiple scatterers that are moving. Even though there are many scatterers in an environment, detection of moving targets is possible using differences of successive radar snapshots. This is generally the case when behind the wall human targets are to be detected. We investigated the effectiveness of various types Multiple Signal Classification (MUSIC) algorithms on the data acquired by our ultra wideband radar prototype. In ideal computer simulations, Time Reversal MUSIC (TRM) algorithm provides successful estimations of both directions and distances of multiple targets. However in practice where non-ideal effects are existent, the performance of TRM algorithm is estimating the target distances degrades. On the other hand, Delay Estimation MUSIC algorithm provides better estimates for the distances of the targets since it is less sensitive to phase noise. Combining the output of TRM algorithm for target directions and the output of Delay Estimation MUSIC method for target distances resulted in successful localization of targets. Experiments are performed using two moving targets in order to test the effectiveness the proposed processing scheme. The problem of detection ambiguities is also considered and several methods to resolve actual targets are presented.