Abstract: A Khatri-Rao subspace based wideband direction-of-arrive (DOA) estimation algorithm for nonlinear arrays without preliminary angle estimation is proposed. From steering vectors of the Khatri-Rao subspace virtual array, the wideband focusing matrix regardless of DOAs is constructed with a manifold separation technique.Benefited from the increased dimensions of the Khatri-Rao subspace virtual array, preliminary angle estimation can be avoided and the algorithm still performs well. On the other hand, by using Root-MUSIC, this method can avoid expensive spectrum searching used in conventional methods so as to reduce the computational burden. Simulations show that performance of the proposed method is close to the preliminary angle estimation needed Khatri-Rao subspace wideband DOA estimation algorithm, FKR-RSS. The proposed method performs better than FKR-RSS when the number of sources is more than the number of sensors.

Key words: array signal processing, direction-of-arrive (DOA) estimation, wideband source, Khatri-Rao subspace, manifold separation technique