Performanceofdirectionofarrival(DOA)estimationissignificantlydegraded inpracticalsituations.Oneoftheconditionsthatdecreasesitsperformance,is the coherent loss caused by the propagation of the wavefront through random non-homogeneous media. Most of the previous methods, such as matrix fitting method and covariance matching technique, need the multidimensional search; therefore their computations are difficult and not suitable for a real-time application. In this paper, a three-stage method based on the generalised eigenvalues utilising signal subspace eigenvectors (GEESE) and Gauss–Newton (GN) algorithms, namely 3SM/GE-GN, is presented. Simulation results show the effectiveness of the proposed method. Difference of RMSE of DOA estimation between proposed method, which is independentofthecoherentloss,andothermethodsismoreclearlyvisiblein higher snapshots. As well, in fixed snapshot, the RMSE of the proposed method is lower than others in different SNRs. Furthermore, the computational time of the simulation is very lower than two previous works

Keywords