Robust target identification in white Gaussian noise for ultra wide-band radar systems

Radar target identification, as witnessed by the plethora of the literature on the topic, is an important problem of considerable interest to many civilian and military agencies. The number of signatures even for a small target library can become quite large since, in general, a unique return is produced for each new target aspect. Any robust target identification algorithm must adequately address this issue. The extinction pulse (E-pulse) and other related techniques, which are based on a singularity expansion method description of the radar return, indeed boast an aspect independent identification algorithm. However, as demonstrated in this paper, the performance of these techniques in white Gaussian noise is inferior to the method described here. In this paper, we develop a new method based on a generalized likelihood ratio test (GLRT) to perform target identification in the presence of white Gaussian noise. As with the E-pulse technique, our method takes advantage of the parsimonious singularity expansion representation of the radar return. In addition, sufficient statistics and simple practical implementations of a GLRT are presented. Simulation results using various thin wire targets are presented contrasting the performance of the GLRT to the E-pulse technique as a function of signal-to-noise (SNR) ratio.