Channel Estimation over Doubly Selective Channels Based on Basis Expansion Model and Compressive Sensing

Abstract

This paper proposes a novel channel estimation (CE) algorithm over doubly-selective (DS) channels based on basis expansion model (BEM) and compressive sensing (CS). In high speed scenarios, the coefficients of DS channel introduce inter-carrier interference (ICI) to orthogonal frequency division multiplexing (OFDM) broadband system; and it needs a large number of pilot subcarriers to estimate the channel matrix. The CE for the system over DS channels is very challenging. Distributed compressive sensing (DCS) theory combined with the BEM is a very efficient way to accurately estimate the channel state information (CSI) of DS channels, but at the overhead cost of the guard pilots around the effective pilots. A novel ICI cancellation based CE algorithm is proposed to estimate the CSI of pilots by eliminating the ICI from the neighborhood subcarriers. The guard pilots around the effective pilots are no longer needed; as results, the number of pilots needed is only 1/5 of the traditional pilot pattern scheme, but at the cost of a limited performance loss. If the same pilot pattern without guard pilots is employed, the proposed algorithm outperforms the traditional CE algorithms.