Constraints implementation for IQML and mode direction-of-arrival estimators

Abstract

The iterative quadratic maximum likelihood IQML and the method of direction estimation MODE are well known high resolution direction-of-arrival DOA estimation methods. Their solutions lead to an optimization problem with constraints. The usual linear constraint presents a poor performance for certain DOA values. This work proposes a new linear constraint applicable to both DOA methods and compare their performance with two others: unit norm and usual linear constraint. It is shown that the proposed alternative performs better than others constraints. The resulting computational complexity is also investigated.