Bi-iteration subspace tracking using instrumental variable

This paper proposes an approximated power method to track the signal subspace. Using rank-one update model, the signal subspace can be obtained by finding the least square solutions of two unconstrained weight functions. As introduced in instrumental variable (IV), this bi-iteration algorithm can track signals corrupted by colored noise. Especially the tracking results are orthonormalized to get better performances. Some DOA simulations are provided to illustrate the effectiveness of the proposed method.