Abstract
In this paper the implementation of the SVD-updating algorithm using orthonormal μ-rotations is presented. An orthonormal μ-rotation is a rotation by an angle of a given set of μ-rotation angles (e.g., the angles Φi = arctan2-i) which are choosen such that the rotation can be implemented by a small amount of shift-add operations. A version of the SVD-updating algorithm is used where all computations are entirely based on the evaluation and application of orthonormal rotations. Therefore, in this form the SVD-updating algorithm is amenable to an implementation using orthonormal μ-rotations, i.e., each rotation executed in the SVD-updating algorithm will be approximated by orthonormal μ-rotations. For all the approximations the same accuracy is used, i.e., onlyr≪w (w: wordlength) orthonormal μ-rotations are used to approximate the exact rotation. The rotation evaluation can also be performed by the execution of μ-rotations such that the complete SVD-updating algorithm can be expressed in terms of orthonormal μ-rotations. Simulations show the efficiency of the SVD-updating algorithm based on orthonormal μ-rotations.
Similar content being viewed by others
References
A. Paulraj, R. Roy, and T. Kailath, “A subspace approach to signal parameter estimation”,Proceedings of the IEEE, Vol. 74, pp. 1044–1045, 1986.
M. Moonen, B. De Moor, L. Vandenberghe, and J. Vandewalle, “On- and off-line identification of linear state-space models”,Int. J. Control, Vol. 49, pp. 219–232, 1989.
S.B. Bensley and B. Aazhang, “Subspace-based channel estimation for code division multiple access communication systems”,IEEE Trans. on Communications (submitted), 1994.
T.F. Chan, “Rank revealing QR factorization”,Linear Algebra and Its Applications, Vol. 89, pp. 67–82, 1987.
G.W. Stewart, “An updating algorithm for subspace tracking”,IEEE Trans. on Signal Processing, Vol. 40, pp. 1535–1541, 1992.
P. Comon and G.H. Golub, “Tracking a few extreme singular values and vectors in signal processing”,Proceedings of the IEEE, Vol. 78, pp. 1327–1343, 1990.
M. Moonen, P. van Dooren, and J. Vandewalle, “A singular value decomposition updating algorithm for subspace tracking”,SIAM J. Matrix Anal. Appl., Vol. 13, pp. 1015–1038, 1992.
W. Ferzali and J.G. Proakis, “Adaptive SVD algorithm for covariance matrix eigenstructure computation”, InProc. IEEE Int. Conf. on Acoustics, Speech, and Signal Processing, Toronto (Canada), 1990, pp. 2615–2618.
M. Moonen, P. van Dooren, and F. Vanpoucke, “On the QR algorithm and updating the SVD and the URV decomposition in parallel”,Linear Algebra and Its Applications, Vol. 188, pp. 549–568, 1993.
F. Vanpoucke, M. Moonen, and E. Deprettere, “A numerically stable Jacobi array for parallel SVD updating”, InSPIE Advanced Signal Processing: Algorithms, Architectures and Implementations V, San Diego (USA), 1994, Vol. 2296, pp. 403–412.
S.F. Hsieh, K.J.R. Liu, and K. Yao, “A unified square-root-free approach for QRD-based recursive least squares estimation”,IEEE Trans. on Signal Processing, Vol. 41, pp. 1405–1409, 1993.
J.E. Volder, “The CORDIC trigonometric computing technique”,IRE Trans. Electronic Computers, Vol. EC 8, pp. 330–334, 1959.
J.S. Walter, “An unified algorithm for elementary functions”, InProc. Spring Joint Computer Conference, AFIPS Press, Vol. 38, p. 397, 1971.
J. Götze, S. Paul, and M. Sauer, “An efficient Jacobi-like algorithm for parallel eigenvalue computation”,IEEE Trans. on Computers, Vol. 42, pp. 1058–1065, 1993.
J. Götze, “An iterative version of the QRD for adaptive RLS filtering”, InProc. SPIE's Int. Conf. on Advanced Signal Processing: Algorithms, Architectures and Implementations, San Diego (USA), 1994.
J. Götze and G.J. Hekstra, “Adaptive approximate rotations for EVD”, In M. Moonen and F. Catthoor, editors,In Algorithms and Parallel VLSI Architectures, Leuven, 1994.
J. Götze and G.J. Hekstra, “An algorithm and architecture based on orthonormal μ-rotations for computing the symmetric EVD”,INTEGRATION—The VLSI Journal. Special Issue on Algorithms and Parallel VLSI Architectures, pp. 21–40, 1995.
J. Götze, “On the parallel implementation of Jacobi and Kogbetliantz algorithms”,SIAM J. on Sci. Comput., Vol. 15, pp. 1331–1348, 1994.
J. Götze, P. Rieder, and J.A. Nossek, “Parallel SVD-updating using approximate rotations”, InSPIE Advanced Signal Processing: Algorithms, Architectures and Implementations VI, San Diego (USA), 1995, pp. 242–252.
J.-M. Delosme, “CORDIC algorithms: Theory and extensions”, InProc. SPIE Advanced Alg. and Arch for Signal Processing IV, Vol. 1152, pp. 131–145, 1989.
B. Yang and J.F. Böhme, “Reducing the computations of the singular value decomposition array given by Brent and Luk”,SIAM J. Matrix Anal. Appl., Vol. 12, pp. 713–725, 1991.
J.R. Cavallaro and F.T. Luk, “CORDIC arithmetic for an SVD processor”,J. Parallel & Distributed Computing, Vol. 5, pp. 271–290, 1988.
J.-M. Delosme, “Bit-level systolic algorithm for the symmetric eigenvalue problem”, InProc. Int. Conf. on Application Specific Array Processors, Princeton (USA), 1990, pp. 771–781.
F.T. Luk and H. Park, “On parallel Jacobi orderings”,SIAM J. Sci. Stat. Comput., Vol. 10, pp. 18–26, 1989.
W.M. Gentleman and H.T. Kung, “Matrix triangularization by systolic arrays”, InSPIE Real-Time Signal Processing IV 298, San Diego (USA), 1981, pp. 19–26.
R.P. Brent and F.T. Luk, “The solution of singular value and symmetric eigenvalue problems on multiprocessor arrays”,SIAM J. Sci. Stat. Comput., Vol. 6, pp. 69–84, 1985.
L. Euler,Opra Omnia I, Vol. 6, pp. 283–315, 1921.
H.W. van Dijk and E.F. Deprettere, “Transformational reasoning on time-adaptive Jacobi-type algorithms”, In3rd Int. Workshop on SVD and Signal Processing, Leuven (Belgium), 1994.
Author information
Authors and Affiliations
Additional information
This work was done while with Rice University, Houston, Texas supported by the Alexander von Humbodt Foundation and Texas Advanced Technology Program.
Rights and permissions
About this article
Cite this article
Götze, J., Rieder, P. SVD-updating using orthonormal μ-rotations. J VLSI Sign Process Syst Sign Image Video Technol 14, 7–17 (1996). https://doi.org/10.1007/BF00925264
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/BF00925264