Data-Assisted Massive MIMO Uplink Transmission with Large Backhaul Cooperation Delay: Scheme Design and System-Level Analysis

It has been shown in the existing literature that data symbols could help to relieve the pilot contamination issue of massive multiple-input multiple-output systems. In this paper, we show that pilot information of interference users could be exploited jointly with the data-assisted transmission mechanism. Specifically, we consider the uplink transmission of a massive multiple-input multiple-output network where there are backhauls with significant delay between base stations. Although real-time inter-base station cooperation is infeasible; pilot information, whose update frequency is very low, of closest interference users can be notified to the serving base station. Conventionally, estimating the interference channel will lead to larger pilot overhead. In our proposed scheme, however, the detected uplink data could provide sufficient degree- of-freedom to estimate the channels of the closest interference users without increasing pilot overhead. Hence, the inter-cell interference can be suppressed efficiently. In order to obtain useful insights on system-level performance, a stochastic geometry based framework is established to analyze the distribution of the signal-to-interference ratio of the proposed scheme. The closed-from expression of the asymptotic bound is thereby derived. It is shown that the analytical expression fits the numerical simulations very well, and the proposed scheme has significant gain over the data-assisted uplink scheme without backhaul.