Robust and reduced-rank matrix wiener filter based on the block conjugate gradient algorithm

Coordinated multi-point transmission (CoMP) allows to deal with interference limitation in cellular systems by acting as a distributed antenna system across cell boundaries. In order to successfully enable CoMP, accurate channel knowledge is required. In our work we focus on uplink joint reception, using linear MMSE combining, suppressing interference. We use LTE-Advanced signal formats together with realistic timevarying frequency-selective channel models. We investigate the performance of different channel estimators in conjunction with several types of Zadoff-Chu-based pilot sequence assignments and different operation points for interference-over thermal noise (IoT) and SIR. The channels are either estimated jointly or separately for all users within the coordinated cell set. We show for Zadoff-Chu sequences that the necessity for joint estimation is much higher for the case of separation by different roots than for the case of separation by different shifts. We also show dependencies of joint estimation gains over IoT and SIR and the interconnection with different approaches for the receive combiner.

show abstract

“…Using the WienerHopf equation, the mean squared error is minimized with the Matrix Wiener Filter (MWF), as adressed e.g. in [8], with weights…”

Section: Joint Lmmse Estimation By a 2d Matrix Wiener Filtermentioning

confidence: 99%