Adaptive algorithms for estimating the complete covariance eigenstructure

This paper presents a new method for adaptively calculating the complete set of eigenvectors and eigenvalues of a recursively updated covariance matrix estimate. The algorithm to be presented is based on a sequence of QR matrix factorisations, currently one of the most efficient methods for computing matrix eigenstructure. A parallel array of O(N²) processing cells is briefly described as a potential architecture for completeting each QR iteration in O(N) time steps.