Superimposed Pilots are Superior for Mitigating Pilot Contamination in Massive MIMO

TL;DR

This paper introduces superimposed pilots as an effective alternative to traditional pilots for reducing pilot contamination in massive MIMO systems, proposing new estimation schemes and hybrid system designs that outperform existing methods.

Contribution

The paper presents a non-iterative channel estimation scheme using superimposed pilots, an iterative data-aided interference reduction method, and demonstrates the superiority of hybrid pilot systems over traditional approaches.

Findings

Hybrid systems outperform pure time-multiplexed pilot systems.

Iterative data-aided schemes improve channel estimation quality.

Superimposed pilots effectively mitigate pilot contamination.

Abstract

In this paper, superimposed pilots are introduced as an alternative to time-multiplexed pilot and data symbols for mitigating pilot contamination in massive multiple-input multiple-output (MIMO) systems. We propose a non-iterative scheme for uplink channel estimation based on superimposed pilots and derive an expression for the uplink signal-to-interference-plus-noise ratio (SINR) at the output of a matched filter employing this channel estimate. Based on this expression, we observe that power control is essential when superimposed pilots are employed. Moreover, the quality of the channel estimate can be improved by reducing the interference that results from transmitting data alongside the pilots, and an intuitive iterative data-aided scheme that reduces this component of interference is also proposed. Approximate expressions for the uplink SINR are provided for the iterative data-aided method as well. In addition, we show that a hybrid system with users utilizing both time-multiplexed and superimposed pilots is superior to an optimally designed system that employs only time-multiplexed pilots, even when the non-iterative channel estimate is used to build the detector and precoder. We also describe a simple approach to implement this hybrid system by minimizing the overall inter and intra-cell interference. Numerical simulations demonstrating the performance of the proposed channel estimation schemes and the superiority of the hybrid system are also provided.