CSIT Estimation and Feedback Reduction for FDD Multi-user Massive MIMO System
Date Issued
2016
Date
2016
Author(s)
Tseng, Po-Ju
Abstract
The concept of massive MIMO is a system where a base station (BS) equipped with a large number of antennas, say 100 antennas or more, serves several users in the same frequency band simultaneously. However, to harvest the benefi ts, BS must acquire instantaneous channel state information (CSI) in order to suppress inter-user interference. By exploiting channel reciprocity under time division duplexing (TDD) operation, CSI acquisition at BS is feasible. Nonetheless, recent researches show that the assumption of channel reciprocity might not be practical. On the other hand, under frequency division duplexing (FDD) operation, channel estimation is much more challenging since resources required by pilot training and feedback overhead both grow linearly with the number of BS antennas. In our work, we exploit the correlation between BS antennas and correlation between UEs due to their close distance. Such correlation means that we can reduce pilot training and feedback overhead signi cantly by using compressive sensing, make CSI acquisition at BS under FDD operation feasible. We further take the scheduling stage into consideration, by combining the order statistics, we can further reduce UEs feedback loads with very little total rate loss.
Subjects
Massive MIMO
Compressive sensing
Order statistics
Type
thesis