Element position considerations for robust direction finding using sparse arrays

Estimating the directions to incoming signals using an array of sensors is of interest in a wide range of applications. A uniform sensor spacing is most often assumed. However, arrays employing a sparse spatial sampling make more efficient use of a given number of sensors. This paper re-examines the use of the Dolph-Chebyshev criterion for selecting the element positions. A simplified implementation is proposed. The array design is linked to estimation performance by a numerical example. It is found that at least what concerns the MUSIC algorithm, a small main-lobe area is more important than low sidelobe levels. This remains true even in scenarios with many signals of different power levels present.