Robust Regularized Block Diagonalization for Mitigating TDD Channel Estimation Errors

Conference: European Wireless 2017 - 23th European Wireless Conference
05/17/2017 - 05/19/2017 at Dresden, Germany

Proceedings: European Wireless 2017

Pages: 7Language: englishTyp: PDF

Personal VDE Members are entitled to a 10% discount on this title

Authors:
Guler, Hilal (Ecole Polytechnique Federale de Lausanne, Information and Communication Sciences, Switzerland)
Bazzi, Samer; Xu, Wen (Huawei Technologies, European Research Center, Riesstr. 25, 80992 Munich, Germany)

Abstract:
In time-division-duplex mode, reciprocity of uplink and downlink channels holds with perfect calibration of the hardware chains at both the base station and user sides. However, hardware calibration errors that exist in practice lead to channel state information (CSI) errors which could result in significant performance degradations in multi-user multiple-input-multipleoutput systems. Further, estimation errors in the uplink constitute another source of CSI errors. To alleviate the degradations due to these errors, we propose a robust regularized block diagonalization (RBD) precoder design to minimize inter-user interference and average bit error rate by making use of calibration, uplink channel estimation errors, and channel statistics. Our analysis properly models calibration and uplink estimation errors as multiplicative and additive errors, respectively. Simulation results demonstrate the gains of the proposed design compared to a nonrobust design. Furthermore, the performance gain of the robust design over the non-robust one increases with increasing signal-to-noise ratios, calibration errors, and number of users. From a practical point of view, a robust RBD precoder can be used to relax calibration requirements.