On local convergence of a class of blind separation algorithms
Ulf Lindgren, Torbjörn Wigren, H. Broman · IEEE Transactions on Signal Processing · 1995
A class of recursive stochastic gradient algorithms for blind separation of dynamically mixed independent source signals are analyzed. The studied methods utilize correlations and high-order moments in order to enforce statistical independence of the separated signals. The local convergence properties of the schemes are investigated, and it is demonstrated that local convergence is tied to positive realness of certain mixing transfer functions.