Abstract
Thomson's multitaper method reduces the variance of power spectral density (PSD) estimates by averaging measurements weighted by orthogonal tapers at the expense of resolution. Co-prime sensing arrays match the resolution of a uniform linear array using fewer sensors by coherently combining interleaved spatially under-sampled subarrays for spatial PSD estimation. This research proposes an algorithm that combines both techniques to compute a spatial PSD estimate with reduced variance with respect to the traditional co-prime sensing array but still uses fewer sensors than a uniform linear array. A multitapered co-prime sensing array reduces variance proportional to the number of tapers used and reduces sensors by at least 33%.