PDF-Euclidean Distance-Based Adaptive Waveform Selection for Maximizing Radar Practical Resolution

This paper focuses on the issue of adaptive waveform selection to optimize the radar resolvability of closely located targets, which is significant in radar detection and estimation. The well-known ambiguity function-based radar resolution only takes transmitted waveform into consideration, which cannot reflect the achievable resolution of a real radar system. In this paper, we consider radar practical resolution based on a geometric metric, Euclidean Distance between probability density functions (PDF-ED), which is defined as square difference between the probability density functions of radar measurements. The PDF-ED takes the PDF's envelope curve into account and specifies the essential difference between the two PDFs in terms of information geometry. Thus the radar practical resolution based on this conveniently characterizes the effect of waveform parameter, target state and measurement model, etc. and offers a statistical way to assess radar sensing capability for a given application. Accordingly, an adaptive waveform selection criterion aiming to maximize the practical resolution is proposed. The experimental simulations verify its effectiveness in decreasing the probability of error when distinguishing two targets.