An iterative multichannel method for blind Volterra system equalization

Abstract

Blind channel equalization is a challenging problem in modern communication systems, where the transmission of a training sequence may not be possible. In this paper a blind multichannel equalizer design procedure is proposed for Volterra (nonlinear) channels. Knowledge of the exact lengths of the system kernels is not necessary, and perfect reconstruction of the input is possible in the noise-free case. Simulations demonstrate the performance of the algorithm on a third-order baseband Volterra system.