Circular to linear array mapping using calibration data

For surveillance and direction finding at HF angular super-resolution techniques can mean significant advantages. Aside from better azimuthal accuracy, elevation resolution and self-calibration are also at hand. Efficient closed form super-resolution algorithms do exist for linear antenna arrays, but unfortunately there are very few for the (needed) circular ones. After giving a short background, this paper describes a least square mapping technique whereby a circular array manifold is mapped onto that of an imaginary linear one so that the existing fast algorithms can be used. The transformation matrix is formed from a relatively dense grid of calibration data. It is found that segmenting the circular array into 30/spl deg/ sectors gives sufficiently small mapping errors, around 0.1/spl deg/. The described technique is currently being implemented in an 8 element 3-30 MHz circular array. The generalized weighted subspace fitting algorithm, GWSF, is used to determine the directions of arrival from the mapped data.