A Single-Snapshot MUSIC Algorithm for ADAS Radar Processing

A single snapshot MUSIC based algorithm is proposed on the framework of ADAS radar processing. It is known that MUSIC is a spectral estimation method that allows to surpass the classical Rayleigh resolution limits. However, the computational cost is much higher than the one required by classical FFT processing, especially if a smoothing procedure has to be employed, as for the single snapshot case. To reduce the numerical complexity here we exploit a Toeplitz data arrangement and split the pseudospectrum formation in two steps. The first one consists of one-dimensional processing whose results are employed as proxy for the multidimensional estimation problem. It is shown that this strategy greatly reduces the cost for building the multidimensional pseudospectrum. Here, for the sake of simplicity, we limit to address the range/angle target estimation.