Performance Evaluation of Beamspace MIMO Systems with Channel Estimation in Realistic Environments

Abstract

Beamspace MIMO (BS-MIMO) systems have been recently proposed as a means to address the two key weaknesses of conventional MIMO systems: the antenna size and the need for multiple RF chains. BS-MIMO transmission is supported by the Electronically Steerable Passive Array Radiators (ESPAR) with a single active and multiple parasitic elements. The main objective is to develop efficient MIMO multiplexing schemes that use only one Radio Frequency (RF) chain and simultaneously maintain extremely small antenna size. Moreover, the recent research results have shown that BS-MIMO systems have increased multiplexing and beamforming capabilities and for small antenna sizes clearly outperform equivalent conventional systems in terms of system capacity. Nevertheless, research on BS-MIMO has been focusing on the study of the ESPAR antenna properties that facilitate beamspace transmission and the theoretical analysis of the ergodic capacity provided by the aerial degrees of freedom (aDoF) of the beamspace channel. Recently, the first steps have been made to design and evaluate practical BS-MIMO systems. This paper presents extended results of the first attempt to design practical and realistic BS-MIMO transmission and reception schemes and it specifically focuses on channel estimation techniques for BS-MIMO systems with adaptive pattern reconfiguration. Adaptation of the basic least-squares (LS) and minimum mean squared error (MMSE) estimators for the beamspace radio channels is performed and the algorithms are incorporated in an adaptive Singular Value Decomposition (SVD)-based system. Finally, fundamental results extracted by the developed beamspace link level simulator are presented in order to evaluate and compare BS-MIMO with equivalent conventional MIMO systems.