Rank Reduced ESPRIT Techniques in the Estimation of Principle Signal Components
J. Andrew Zhang, Rodney A. Kennedy, Thushara Dheemantha Abhayapala · 2004
In this paper, we present novel rank reduced ESPRIT algorithms to estimate principle signal components with low computational complexity, which can typically be applied in the high resolution identification of closely spaced wireless multipath channels. These algorithms transform the generalized eigen-problem from an original high dimensional space to a lower dimensional space depending on the number of desired principle signals. As only principle singular values and vectors are required, fast algorithms such as the power method can be applied to greatly simplify the proposed algorithm and make it implementable in real time.