The　performance　of　the　subspace-based　direction　of　arrival　(DOA)　estimation　degrades　in　coherent　source　scenarios　due　to　rank　deficiency　of　the　covariance　matrix　of　array-received　signals.　To　solve　this　problem,　various　methods　such　as　the　spatial　smoothing　pre-processing　(SSP)　and　covariance　matrix　reconstruction　methods　have　been　proposed.　However,　these　methods　do　not　make　fully　use　of　the　available　information　in　the　covariance　matrix　and　their　performance　deteriorates　under　low　signal-to-noise　ratio　(SNR)　conditions.　In　this　paper,　we　propose　a　DOA　estimation　method　for　coherent　sources　which　is　robust　to　low　SNR.　Specifically,　a　Hankel　matrix　is　constructed　by　the　elements　in　the　rank-1　signal　subspace　of　the　covariance　matrix　of　the　denoised　array-received　signals,　in　order　to　recover　a　rank-k　signal　subspace.　Then,　the　MUSIC　method　is　used　with　the　noise　subspace　of　this　Hankel　matrix　for　coherent　source　localization.　The　simulation　results　show　that　the　proposed　algorithm　can　make　full　use　of　the　available　information　in　the　covariance　matrix　of　array-received　signals　and　effectively　estimate　the　DOA　of　coherent　sources　under　low　SNR　conditions.　©2024　IEEE.