摘要： An e?cient algorithm based on high-order cumulant is addressed for the scenarios where both far-?eld and near-?eld narrow-band signals may exist synchronously. The ?rst matrix built by four-order cumulant is utilized to estimate the two dimensional direction-of-arrivals (DOAs) using the orthogonal projection matrix of the signal subspace and the virtual steering matrix. Then, the second matrix built by four-order cumulant is decomposed to get the noise subspace using the eigen decomposition. Meanwhile, a virtual steering matrix is used to distinguish far-?eld signals (FFSs) from near-?eld signals (NFSs). And one-dimensional MUSIC algorithm is used to estimate the range of the NFSs. Compared to the TSMUSIC, the proposed algorithm can provide high resolution for the DOAs. In addition, there is higher accuracy for the DOA of NFS in the proposed algorithm than that in TSMUSIC and in TSMD. Simulation results are carried out to certify the performance of the proposed algorithm.