Covariance-Based DoA Estimation in a Krylov Subspace

Covariance-based DoA estimation (CB-DoA) algorithms represent lower computational complexity alternatives to the traditional ESPRIT approach. This paper investigates CB-DoA using Krylov-subspace techniques (including Arnoldi's and Lanczos' updates) with respect to the resulting computational cost and estimation error performance. The proposed modifications also allow an automatic estimation of the number of sources. Computational analyses performed for the resulting CB-DoA algorithm indicate cost savings above in comparison with the standard CB-DoA implementation, which already represents a improvement upon its ESPRIT counterpart, at an equivalent mean squared error level, as verified in numerical simulations.