No abstract

The objective of this paper is to study the performance of a multicell vector perturbation (MVP) precoding technique under practical situations in a network multiple-input multiple-output (MIMO) scheme employing joint transmission. The conventional perturbation strategy that minimizes the total power is considered, and the power at each base station (BS) is properly scaled to enforce per-BS power constraints. In our scenario, we consider multiple-antenna users and use block diagonalization (BD) as the linear front-end precoder for VP. The sum rate for the MVP in the case of uniformly distributed input and an asymptotic upper bound on the sum rate at high signal-to-noise ratios (SNRs) are derived. In addition, using the asymptotic upper bound on the individual user rates, we propose a proportionally fair (PF) user scheduling algorithm of lower complexity and better performance compared with the benchmark fair semiorthogonal user selection (SUS) algorithm. As opposed to the PF-SUS, the proposed PF scheduling algorithm requires no predefined correlation threshold. Furthermore, we study the impact of backhaul delay on the performance of both VP and BD by deriving bounds on the sum rate. The numerical results show that MVP in the case of perfect channel state information (CSI) outperforms multicell BD. In the presence of a backhaul delay, the performance of MVP significantly degrades, but the upper bound on the sum rate for MVP is still higher than for multicell BD.Index Terms-Backhaul delay, cellular downlink, joint transmission, network multiple-input multiple-output (MIMO), proportionally fair (PF) user scheduling, vector perturbation (VP) precoding.

In this paper, we consider time domain vector perturbation (TDVP) in a large-system limit when channel state information (CSI) is imperfect due to pilot contamination and we derive the system's achievable sum rate per cell. We use random matrix methods to avoid time-consuming MonteCarlo simulations. Numerical results show that in general linear precoding ensures higher sum rate than time domain vector perturbation in the massive MIMO regime. Index Terms-Massive MIMO, large-system analysis, time domain vector perturbation (TDVP). I. INTRODUCTIONIn a Gaussian MIMO broadcast channel it has been proved that dirty paper coding (DPC) [1]-[3], which is a very complex non-linear precoding technique, delivers the entire capacity region and achieves the sum capacity [2]-[4]. There exist other non-linear techniques of lower complexity but suboptimal, such as Tomlinson-Harashima precoding (THP) [5], [6] and vector perturbation (VP) [7]. Analysis of non-linear techniques employing VP is much more complicated than that of linear techniques because perturbing vectors are highly correlated and data-dependent. Adding to that, imperfect CSI and user scheduling make the analysis even harder [8]. There are several publications, which attempt to examine analytically the behavior of VP techniques [9]-[11]. However, these works assume that CSI is perfectly available at the transmitter or otherwise give upper and lower bounds on the sum rate. Avner et al. [12] proposed a VP technique operating in time domain, called time domain VP (TDVP), where the data vector for each user is perturbed in time domain instead of user domain. In this paper, we investigate the performance of TDVP in a largesize system. We use the framework developed in [13]-[15] and random matrix theory [16] to avoid time-consuming MonteCarlo simulations and get better insight into the problem.

Recently, studies on sub-optimal precoding techniques for multiple-input multiple-output broadcast channels (MIMO-BC), which achieve performance near to that of the dirty paper coding (DPC), have drawn attention to vector perturbation (VP). In practice, each antenna or more generally each antenna group has its own limit on the transmitted power, which makes per-antenna-group (per-AG) power constraints more meaningful than the sum power constraint. In this paper, we assume the perantenna-group constraints and first investigate antenna-group power-constrained VP inspired by the p-sphere encoder for the case when channel inversion is used by the front-end linear precoder. Furthermore, we consider joint optimization of the front-end precoder and the perturbing vector subject to per-AG power constraints and propose precoding based on a minimum mean square error (MMSE) criterion. The results show that the proposed algorithm outperforms conventional VP and the p-sphere encoding in the case of per-AG power constraints.

In this paper, a practical approach to pilot contamination precoding (PCP) for massive MIMO is proposed through a joint clustering and pilot reuse scheme. We also introduce power scaling to enforce per-base station (BS) power constraints. We consider a massive MIMO system, where uncoordinated conventional beamforming is implemented in each cell. PCP acts as outer linear precoding prior to conventional beamforming through a cooperative transmission scheme with 3 base stations (BSs) involved. We partition each cell into 3 sectors and assign pilot sequences in a suitable way in order to perform PCP. In order to characterize performance and avoid time-consuming simulations, we employ large system analysis and random matrix theory. Numerical results show that the superiority of the clustered PCP is marginal for the moderate number of transmit antennas, but it becomes more significant in a massive MIMO mode. In addition, depending on user location, some users may experience a two-fold increased spectral efficiency after applying clustered PCP in the massive MIMO mode.