A new algorithm for computing the extreme eigenvectors of a complex Hermitian matrix

This paper presents a novel algorithm for computing the eigenvector associated with either the largest or the smallest eigenvalue of a complex Hermitian matrix. This type of algorithm is required for direction of arrival (DOA) and frequency estimation. Necessary and sufficient conditions for convergence are proved, and simulations show the superior performance over traditional methods.

[1]  Jonathan H. Manton,et al.  Modified steepest descent and Newton algorithms for orthogonally constrained optimisation. Part I. The complex Stiefel manifold , 2001, Proceedings of the Sixth International Symposium on Signal Processing and its Applications (Cat.No.01EX467).

[2]  Sun-Yuan Kung,et al.  On gradient adaptation with unit-norm constraints , 2000, IEEE Trans. Signal Process..

[3]  J.H. Manton.  An improved least squares blind channel identification algorithm for linearly and affinely precoded communication systems , 2002, IEEE Signal Processing Letters.

[4]  John B. Moore,et al.  Gradient algorithms for principal component analysis , 1996, The Journal of the Australian Mathematical Society. Series B. Applied Mathematics.

[5]  Y. Hua,et al.  Precoder assisted channel estimation in complex projective space , 2001, 2001 IEEE Third Workshop on Signal Processing Advances in Wireless Communications (SPAWC'01). Workshop Proceedings (Cat. No.01EX471).

[6]  G. Golub,et al.  Tracking a few extreme singular values and vectors in signal processing , 1990, Proc. IEEE.

[7]  Elijah Polak,et al.  Optimization: Algorithms and Consistent Approximations , 1997 .

[8]  Jar-Ferr Yang,et al.  Adaptive eigensubspace algorithms for direction or frequency estimation and tracking , 1988, IEEE Trans. Acoust. Speech Signal Process..

[9]  Gene H. Golub,et al.  Matrix computations , 1983 .