Utilizing the reduced-rank technique of the multi-state wiener filter(MSWF), a novel method of direction of arrival(DOA) estimation is proposed, resulting in a new way for parameter estimation. Analysis and simulation show that this method can be applied in the case where only a few samples are available and the signal of interference plus noise ratio(SINR) is very low. On the other hand, the new method requires much lower computational cost than the subspace-based methods. Finally, simlations are given to validate the method.
