Subspace-based direction finding in partly calibrated arrays of arbitrary geometry

The problem of direction finding in partly calibrated arrays of an arbitrary geometry is addressed. We assume that an array is composed of multiple calibrated subarrays and consider the cases of unknown (or known with some error) inter-subarray displacements, imperfect synchronization of subarrays in time, as well as unknown channel mismatches between subarrays. Recently, the so-called RAnk Reduction Estimator (RARE) has been proposed as a particular solution to the problem of direction finding in partly calibrated arrays with unknown inter-subarray displacements. However, the application of RARE is restricted by the special case of arrays composed of identically oriented linear subarrays with interelement spacings which are integer multiples of a certain shortest baseline. In this paper, we propose a more general subspacebased approach which further develops the ideas of RARE and is applicable to partly calibrated arrays composed of subarrays of arbitrary geometry. The algorithms proposed enable simple extensions to the two-dimensional (2D) Direction-Of-Arrival (DOA) estimation case.