IEEE International Mediterranean Conference on Communications and Networking (MeditCom)
Extending Quantum-Annealing-Aided Multi-User Detection to QAM Signals
Kouki Yonaga, Soma Hayabe, Kenichi Takizawa
Abstract
This study presents a quantum-annealing-aided multi-user detection (QA-aided MUD) framework for quadrature amplitude modulation (QAM) signals and evaluates its performance. The proposed approach approximates log-likelihood ratios (LLRs) using quantum annealing, enabling accurate detection without exhaustive computation. Previous studies have primarily focused on quadrature phase shift keying (QPSK) modulation and have shown that QA-aided MUD can outperform conventional low-complexity detectors, particularly in overloaded multi-user multi-input multi-output (MU-MIMO) systems in uplink. In this study, we extend the QA-aided MUD framework to support QAM modulation by constructing a binary-variable formulation and incorporating a bit-level conversion scheme to handle Gray-coded constellations. For performance evaluation, block error rate (BLER) analysis is conducted using a digital annealing, while small-scale bit error rate (BER) analysis is performed using the D-Wave quantum annealer. The results demonstrate competitive, and often superior, BLER performance compared with conventional low-complexity detectors, particularly in overloaded scenarios. However, the detection performance becomes increasingly sensitive to annealing parameters as the modulation order increases. Furthermore, BER analysis indicates that the current D-Wave hardware may fail to reliably detect QAM symbols in some large problem instances. Based on these findings, future challenges for QA-aided MUD are discussed.