Learning Sequential Channel Selection for Interference Alignment using Reconfigurable Antennas

TL;DR

This paper introduces a sequential learning approach for reconfigurable antennas to optimize interference alignment in multi-user wireless networks, reducing the need for full channel state information and improving network sum rate.

Contribution

It formulates the channel selection as a multi-armed bandit problem and demonstrates that adaptive sequential learning can enhance interference alignment without extensive CSI.

Findings

Improved sum rate through learned channel selection

Effective interference alignment with limited CSI

Quantified benefits of pattern diversity

Abstract

In recent years, machine learning techniques have been explored to support, enhance or augment wireless systems especially at the physical layer of the protocol stack. Traditional ML based approach or optimization is often not suitable due to algorithmic complexity, reliance on existing training data and/or due to distributed setting. In this paper, we formulate a reconfigurable antenna based channel selection problem for interference alignment in a multi-user wireless network as a learning problem. More specifically, we propose that by using sequential learning, an effective channel or combination of channels can be selected in order to enhance interference alignment using reconfigurable antennas. We first formulate the channel selection as a multi-armed problem that aims to optimize the sum rate of the network. We show that by using an adaptive sequential learning policy, each node in the network can learn to select optimal channels without requiring full and instantaneous CSI for all the available antenna states. We conduct performance analysis of our technique for a MIMO interference channel using a conventional IA scheme and quantify the benefits of pattern diversity and learning channel selection.