Reducing Feedback Requirements of the Multiple Weight Opportunistic Beamforming Scheme via Selective Multiuser Diversity

Abstract

Opportunistic beamforming (OB) relies on the transmission of Channel State Information (CSI) in the form of instantaneous Signal to Noise Ratio (SNR) from Mobile Stations (MSs) back to the Base Station (BS) for scheduling purposes that increase throughput and/or maintain resource allocation fairness. OB is employed in environments of low mobility and low scatter, artificially inducing channel fluctuations that can better exploit Multiuser diversity (MUD). Multiple antennas at the BS are randomly alter the channel's response and generate peaks in gain where users can be scheduled on, maximising system throughput. Additional gains are achieved by transmitting multiple weighting vectors from the BS, but their use can significantly increase the load of the feedback channel and mitigate MUD gains. Selective Multiuser diversity (SMUD) has been proposed for the original beamforming scheme as a technique that reduces feedback requirements substantially without any significant throughput degradations. This paper considers the joined use of multiple weights and SMUD in opportunistic beamforming, aiming to increase capacity while reducing feedback overhead. Results show that with an appropriate use of threshold levels, not only average throughput can be increased through the use of multiple weighting vectors but also a notable decrease in feedback load requirements compared to the conventional OB with SMUD design is achieved. The terms Mobile Station (MS) and User are used interchangeably in this paper, with the same meaning.

Opportunistic transmission in MU-MIMO downlink with reduced feedback and mode adaptation

“…For simplicity, 1 th Q and 2 th Q are of the same value. Using numerical searching we obtain the threshold values under N T =8 and N k =2 in Table I, analogous to [12]. Given target capacity or outage probability and the statistical characteristic of ζ i , the threshold could also be determined.…”

“…Assume each right singular vector v k,i (t) is quantized into M bits. The feedback load of L eligible users, of which l 1 users select BF at slot t is given by (12). Here CQI feedback load is not considered.…”

“…Other works focus on opportunistic beamforming (OB) [8][9][10][11][12], i.e. at each time slot only one user among multiple candidates is selected.…”

“…Furthermore, when OB is under discussion CCI is no longer considered. Among the works dealing with OB [8,9] investigate the strategy to achieve multiuser diversity with low feedback cost, [10,11] discuss the way to induce large and fast channel fluctuations and [12] takes both issues into account. In these works each MS is equipped with only one receive antenna, i.e.…”

Performance Analysis for Partial Feedback Downlink MIMO with Unitary Codebook Beamforming for LTE

“…Therefore the combined use of OFDMA and multiple precoding design results in an excessive feedback overhead. For this reason various approaches for uplink feedback reduction in OFDMA and multiple precoding selection have been developed [12,13]. This section outlines the basic procedure followed for the unitary beamforming, illustrated in Figure 1, based on a pre-defined codebook with partial feedback design, as proposed in [3].…”