Learning-based Rate Adaptation for Uplink Massive MIMO Networks with Cooperative Data-Assisted Detection

TL;DR

This paper introduces a cooperative uplink transmission scheme for massive MIMO networks that adapts data rates based on interference statistics without prior knowledge of user density, improving reliability.

Contribution

It proposes a novel cooperative detection scheme and an online rate adaptation algorithm that does not require prior user density information.

Findings

Derived the asymptotic SINR for the scheme.

Proved interference power distribution is Gaussian.

Developed a robust rate adaptation algorithm.

Abstract

In this paper, the uplink adaptation for massive multiple-input-multiple-output (MIMO) networks without the knowledge of user density is considered. Specifically, a novel cooperative uplink transmission and detection scheme is first proposed for massive MIMO networks, where each uplink frame is divided into a number of data blocks with independent coding schemes and the following blocks are decoded based on previously detected data blocks in both service and neighboring cells. The asymptotic signal-to-interference-plus-noise ratio (SINR) of the proposed scheme is then derived, and the distribution of interference power considering the randomness of the users' locations is proved to be Gaussian. By tracking the mean and variance of interference power, an online robust rate adaptation algorithm ensuring a target packet outage probability is proposed for the scenario where the interfering channel and the user density are unknown.