Low-complexity initialization of adaptive equalizers using approximate channel inverse

Abstract

In this paper, a low-complexity approximate channel inverse initialization scheme for blind equalization is proposed. The underlying idea is that the inverse of minimum phase finite impulse response (FIR) channels can be well approximated by inverting the most significant paths of the estimated channel impulse response (CIR). The parameters of the inverse of the truncated channel can be expressed in closed-form expressions in terms of the original CIR and can be used to initialize adaptive equalizers. By applying time-reversal method, the proposed initialization scheme can also be applied for equalization of maximum phase channels. Through extensive computer simulations, we show that channels that cannot be equalized by blind equalizers with the conventional single-spike initialization can now be well equalized with the proposed initialization