Estimation of DOA in unknown noise: performance analysis of UN-MUSIC and UN-CLE, and the optimality of CCD

In a previous paper, a new approach was proposed for the consistent estimation of the directions of arrival (DOA) of signals in an unknown spatially correlated noise environment using generalized correlation decomposition (GCD). Based on the various interesting properties of the eigenspace structure obtained by GCD, two effective methods (UN-MUSIC and UNCLE) of estimating the DOA in an unknown correlated noise were developed, In this paper, the performance of the two methods are analyzed. It is shown that the performance of these two methods can be optimized by assigning optimum weighting matrices in their respective criteria. Furthermore, and more importantly, it is also shown that of all the correlation decompositions, the canonical correlation decomposition (CCD) leads to the optimum performance of the methods. Computer simulations confirm these conclusions and show that the use of CCD is robust even under variable spatially correlated noise conditions.<>