Rydberg Atomic Quantum Receivers for Multi-Target DOA Estimation
Tierui Gong, Chau Yuen, Chong Meng Samson See, Mérouane Debbah, Lajos Hanzo
TL;DR
This work develops a RAQR-based uniform linear array model for multi-target DOA estimation and identifies LO-induced sensor gain mismatch as a fundamental challenge for classical ESPRIT. It introduces RAQ-ESPRIT, a DOA estimator that compensates LO-phase effects via a targeted subspace approach, and demonstrates through simulations that RAQ-ESPRIT can achieve massive improvements in estimation accuracy over traditional RF receivers, especially in photon-shot-noise and standard quantum limit regimes. The results highlight the practical potential of quantum RF receivers for high-precision sensing and the viability of RAQ-ESPRIT as a low-complexity, high-performance solution. This work thus paves the way for quantum-domain multi-target sensing with compact RAQR hardware.
Abstract
Quantum sensing technologies have experienced rapid progresses since entering the `second quantum revolution'. Among various candidates, schemes relying on Rydberg atoms exhibit compelling advantages for detecting radio frequency signals. Based on this, Rydberg atomic quantum receivers (RAQRs) have emerged as a promising solution to classical wireless communication and sensing. To harness the advantages and exploit the potential of RAQRs in wireless sensing, we investigate the realization of the direction of arrival (DOA) estimation by RAQRs. Specifically, we first conceive a Rydberg atomic quantum uniform linear array (RAQ-ULA) aided wireless receiver for multi-target DOA detection and propose the corresponding signal model of this sensing system. Our model reveals that the presence of the radio-frequency local oscillator in the RAQ-ULA creates sensor gain mismatches, which degrade the DOA estimation significantly by employing the classical Estimation of Signal Parameters via Rotational Invariant Techniques (ESPRIT). To solve this sensor gain mismatch problem, we propose the Rydberg atomic quantum ESPRIT (RAQ-ESPRIT) relying on our model. Lastly, we characterize our scheme through numerical simulations, where the results exhibit that it is capable of reducing the estimation error of its classical counterpart on the order of $> 400$-fold and $> 9000$-fold in the PSL and SQL, respectively.
