Semiblind bussgang equalization for sparse channels

In this correspondence, we describe a semiblind Bussgang equalization scheme that incorporates the training sequence in a quite direct and flexible fashion. Among others, we discuss how the training sequence can be used to control the Bussgang nonlinearity thus enforcing the convergence of the iterative equalizer design. Moreover, we focus on the case of channels whose impulse response is long but sparse, counteracted by equalizers designed on a suitable sparse support. Numerical performance analysis conducted on compact and sparse time-varying channels shows that sparse Bussgang semiblind equalizer using a short training sequence outperforms blind, trained, and decision directed equalizers in a variety of fading conditions.