Sparse channel estimation based on compressed sensing for ultra wideband systems

Abstract

Channel estimation for purposes of equalization is a long standing problem in signal processing. Wireless propagation is characterized by sparse channels, that is channels whose time domain impulse response consists of few dominant multipath fingers. This paper examines the use of Compressed Sensing (CS) in the estimation of highly sparse channels. In particular, a new channel sparse model for ultra-wideband (UWB) communication systems based on the frequency domain signal model is presented. A new greedy algorithm named extended OMP (eOMP) is proposed to reduce the false path detection achieved with classical Orthogonal Matching Pursuit (OMP) allowing better time of arrival (TOA) estimation.