Least-Squares Based Direction-of-Arrival Estimation using Sparse Circular Arrays

Accurate, least-squares based bias free direction-of-arrival (DOA) estimation from targets using a circular array is discussed. The proposed estimation method is based on a recently proposed decorrelation technique that uses regression sum of squares. The estimation meets Cramer-Rao bounds for any circular array geometry that uses either a large or a small number of sensors. The method is versatile as it can be used in accurate angle estimation even when sensors are placed sparsely and non-uniformly on the circle. The method is computationally efficient since matrix inversions can be avoided. Possibility of DOA estimation from extremely close-by targets with very small angles of separation is demonstrated.