Quantum Annealing-Aided Multi-User Detection: An Application to Uplink Non-Orthogonal Multiple Access
Kouki Yonaga, Kenichi Takizawa · 2023
A new method to realize computation of the log-likelihood ratio (LLR) in multi-user detection (MUD) by quantum annealing (QA) is proposed herein. One metaheuristic algorithm for solving combinatorial optimization problems is QA. Recently, QA has been applied in various fields to hasten computations. This study specifically examines the application of QA in up-link non-orthogonal multiple access (UL-NOMA) that employs iterative MUD. Our proposed method, QA-aided MUD, uses QA to calculate LLR in iterative MUD. We tested QA-aided MUD with a D-Wave quantum annealer. The error rate analysis demonstrates that QA-aided MUD has comparable performance to that of conventional MUD method. Furthermore, when the annealing time was set to 20 µs, we demonstrated than that the total computation time is 1/10 of the computation time necessary when using a digital computer. We also evaluated the convergence behavior of QA-aided MUD through evolution of mutual information. The results indicate that QA-assisted MUD has almost identical convergence performance to that of the conventional MUD method for LLR calculations.