A Recursive Blind Adaptive Identification Algorithm and Its Almost Sure Convergence
Miloje S. Radenković, Tamal Bose · IEEE Transactions on Circuits and Systems I Fundamental Theory and Applications · 2007
This paper presents a novel blind adaptive identification algorithm based on least-squares type arguments. Parameter estimates are recursively updated with each output measurement, without resorting to any matrix inversion operation. It is proved that the parameter estimates converge almost surely (a.s.) toward a scalar multiple of the true parameters. Possible application of this algorithm to the channel equalization problem is discussed.