Multiuser/MIMO Doubly Selective Fading Channel Estimation using TM training and CE-BEM

Abstract

Wireless Communications at high data rates over high mobility platforms are likely to experience frequency and time selective (doubly selective) fading. This paper deals with an efficient estimation and corresponding equalization on doubly-selective channels by using both complex exponential basis expansion model (CE-BEM) and time multiplexed (TM) training under multiuser/multi input multi output (MIMO) scenario. The time domain variations of the doubly selective channel is described by CE-BEMs so that the estimation of a time varying process qualifies for reduction to calculating a few time invariant BEM coefficients. In case of channels that are time-varying, one has to frequently and periodically send training sequences for keeping track of the modifications in the channel. The basis coefficients are obtained via the proposed minimum mean square error estimation (MMSE) technique, which outperforms the conventional least squares (LS) estimation technique. After obtaining the channel estimates at the receiver, the transmitted symbols are detected by using maximum likelihood (ML) detection. In order to achieve high data rates and better system performance we extend the single input single output (SISO) system to multiuser/multi input multi output (MIMO). The performance improvements involved in this proposed approach are obtained through simulation studies.