Robust Detection for Forward-Looking GPR in Rough-Surface Clutter Environments

We propose a robust likelihood ratio test to detect shallowly buried targets in forward-looking ground penetrating radar. Instead of modeling the distributions via parametric families, we consider an uncertainty model by constructing a band of feasible probability densities under each hypothesis. The likelihood ratio test is then designed based on the least favorable densities within the bands. This detector has been shown to maximize the worst-case performance over all feasible density pairs and, hence, avoids strong assumptions about the clutter and noise distributions. Its performance is evaluated numerically and compared to alternative approaches in the literature.