Interference-Aware Scheduling in the Synchronous Cellular Multi-Antenna Downlink

Abstract

Abstract —We consider the downlink of a multi-user MIMOOFDM system in a cellular environment, where users arescheduled to the transmission resources in frequency and space.Targeting a practical solution, we assume the use of fixed DFT-based pre-coding beams at the base stations and linear receiversat the multi-antenna terminals. After having received feedback onthe frequency-selective interference conditions from all terminals,each base station schedules the terminals to their resources ofhighest quality while following a fair scheduling policy. It isshown that this rather simple practical concept enables spatialmultiplexing transmission for all users in the system, in particulareven for users at cell edge. This results in a capacity scaling that issimilar to the one known for isolated point-to-point links. Finally,the system is shown to achieve a performance close to the idealcase if proper techniques to estimate the desired and interferingchannels are applied. I. I NTRODUCTION Transmission with multiple antennas both at the transmittingand receiving ends of a wireless link has become increasinglymature in recent years. From theory, the fundamental capacitygain in the multiple-input multiple-output (MIMO) radio link,being proportional to the minimum of the number of transmitand receive antennas, is well understood for an isolated point-to-point link. A fundamental trade-off between different trans-mission modes has been pointed out in [1], which are spatialmultiplexing (SMUX) and spatial diversity (SDIV), revealingthat the mode maximizing the capacity depends on the actualchannel state. This fact motivated the development of anadaptive transmission system, selecting the transmission modedepending on the actual channel quality in order to improvethe error rate performance for fixed data rate transmission [2]or to increase the spectral efficiency [3].To enable ubiquitous broadband wireless access, MIMOtransmission must be made robust against multi-cell inter-ference. However, it has not been fully evident yet how thepotential capacity gains of MIMO can be realized under theseconditions. In fact, early results obtained for a small set oflinear transceiver setups, indicate only small gains for SMUXover SDIV systems [4]. The achievable spectral efficiencymay be enhanced by enabling multi-user MIMO (MU-MIMO)into system design and thus turning the focus to multi-userlinks [5]. However, base stations (BSs) would require coherentchannel state information to optimally serve their users inMU-MIMO, which is difficult to obtain in frequency division