Constraints for linear predictive and minimum-norm methods in bearing estimation

Abstract

It is desirable to reduce the computational complexity and improve the resolution of an estimator whenever possible. It is shown in the paper that, for a spatial spectrum estimator using either the linear predictive method or the minimum-norm method, the number of constraints can be reduced by one when the optimum weight vector is obtained using real vector notation. The reduction in the number of constraints in the estimation implies a reduction in computational complexity and a possibility of an increase in the degree of freedom in solving the optimisation problem. This also implies that the modified spectral estimate should be as good, if not better, than the original spectral estimate in terms of resolution performance.