System level simulation and optimization of multi-user MIMO transmissions

Abstract

Providing higher data rates in mobile communication systems without increasing bandwidth can be achieved with techniques such as multi-user multiple-input and multiple-output transmissions using linear precoding. One popular precoding strategy that is often used in practice is zero-forcing precoding, which utilises multiple antennas to cancel the interference between users. The performance of zero-forcing precoding strongly depends on the similarities of the channels of scheduled users. Therefore, a user grouping heuristic is proposed that optimises the transmission in terms of sum-through put. In addition, an extension is introduced to trade-off through put against fairness among users. The proposed algorithm is compared to a semi-orthogonal user selection algorithm by performing system level simulations. The performance of the algorithms is also investigated in a cellular scenario together with fractional frequency reuse.