Multidimensional spectral analysis for radar target parameter estimation: Survey and benchmarks

Most state-of-the-art radar processors are based on sequential, one-dimensional signal processing stages. In this paper we investigate the suitability of modern multidimensional spectral analysis methods for the estimation of radar target parameters. A short survey of both classical and more advanced high- and superresolution multidimensional spectrum estimation techniques like 2D-AR, 2D-Capon, and 2D-MUSIC is given. The performance of these techniques is evaluated by systematic benchmarks, including sensitivity towards noise, spectral resolution, high signal dynamics, and multi-tone capacity. The benchmark results are assessed with regard to joint spatial and temporal radar parameter estimation.