Fast retrieval of signal subspace of over-determined low-rank data matrix

To investigate fast retrieval of signal subspace in signal processing applications in reference to a kind of over-determined low-rank data matrix, the paper makes use of the relationship between left and right singular vectors of singular value decomposition, the anti-power method and L-D-L^T decomposition, and presents a fast algorithm in the retrieval of signal subspace. It obviously involues less computational complexity than is singular value decomposition, attains higher precision, and is suitable for corresponding signal processing.