Improving the Robustness of the Rare Algorithm Against Subarray Orientation Errors

We study the problem of direction-of-arrival (DOA) estimation using partly calibrated arrays composed of multiple subarrays with unknown inter-subarray parameters and imperfectly known subarray orientations. The recently developed spectral and root variants of the rank reduction estimator (RARE) can handle scenarios where no calibration between subarrays is available but, un-fortunately, they are very sensitive to subarray orientation errors. Therefore, conventional RARE can be applied to such partly calibrated arrays only if all subarray misorientations are negligibly small. In this paper, we develop a new modification of RARE which improves its robustness against subarray misorientations. The performance of the proposed robust RARE algorithm is demonstrated to be close to the stochastic Cramer-Rae bound (CRB) of the considered estimation problem.