A finite-interval constant modulus algorithm
Phillip A. Regalia · IEEE International Conference on Acoustics Speech and Signal Processing · 2002
A finite-interval constant modulus algorithm is developed which is vastly simpler than the Analytic Constant Modulus Algorithm and, unlike that algorithm, can claim to minimize a constant modulus criterion. It requires one QR decomposition of a data matrix, followed by a power iteration. Step size bounds which ensure monotonic convergence are obtained in analytic form, and proper tuning leads to an algorithm which converges typically within a few iterations. The algorithm thus gives a computationally feasible method for implementing constant modulus signal restoration in packet-based transmission systems.