X-ResQ: Reverse Annealing for Quantum MIMO Detection with Flexible Parallelism

TL;DR

X-ResQ tackles the challenge of near-optimal MIMO detection under tight latency by leveraging Reverse Annealing-based parallel QA. It introduces a multi-seed ensemble RA framework and a split-detection mechanism to mitigate high-SNR BER floors, validated on D-Wave hardware and via classical PIC simulations. The results show substantial BER/throughput gains over prior QA detectors and demonstrate the potential to scale to ultra-large MIMO in classical contexts. The work discusses practical integration considerations for C-RAN and outlines avenues toward end-to-end quantum-assisted wireless systems.

Abstract

Quantum Annealing (QA)-accelerated MIMO detection is an emerging research approach in the context of NextG wireless networks. The opportunity is to enable large MIMO systems and thus improve wireless performance. The approach aims to leverage QA to expedite the computation required for theoretically optimal but computationally-demanding Maximum Likelihood detection to overcome the limitations of the currently deployed linear detectors. This paper presents X-ResQ, a QA-based MIMO detector system featuring fine-grained quantum task parallelism that is uniquely enabled by the Reverse Annealing (RA) protocol. Unlike prior designs, X-ResQ has many desirable system properties for a parallel QA detector and has effectively improved detection performance as more qubits are assigned. In our evaluations on a state-of-the-art quantum annealer, fully parallel X-ResQ achieves near-optimal throughput (over 10 bits/s/Hz) for $4\times6$ MIMO with 16-QAM using six levels of parallelism with 240 qubits and $220~μ$s QA compute time, achieving 2.5--5$\times$ gains compared against other tested detectors. For more comprehensive evaluations, we implement and evaluate X-ResQ in the non-quantum digital setting. This non-quantum X-ResQ demonstration showcases the potential to realize ultra-large $1024\times1024$ MIMO, significantly outperforming other MIMO detectors, including the state-of-the-art RA detector classically implemented in the same way.

Figures (20)

• Figure 1: MIMO model and optimal ML solution.
• Figure 2: Optimization search of Reverse Annealing (RA) initiated from a classical state (cf. from a superposition state in FA).
• Figure 3: Machine QPU operation for QA and parallel QA.
• Figure 4: System Architecture of X-ResQ (multi-seed parallel ensemble RA). X-ResQ is based on the currently-deployed MMSE detector, requiring only simple preprocessing. Unlike IoT-ResQ, X-ResQ can support flexible parallelism with fine granularity for any $L_P$, where all the tasks can converge to the ML solution, thus increasing the ML fidelity through sample parallelism.
• Figure 5: TTS Analysis of RA that is initialized from H. Dist=1 initial states using a $4\times4$ MIMO detection instance ($N_V=16$) at SNR 20 dB. It demonstrates initial states that have lower Ising energies do not necessarily result in better RA results.