HQC2NN: Hybrid Quantum-Classical Drone Detection for Low-SNR Conditions in Low-Altitude Economy Networks

Abstract

Recently, shortage of ground-based communication, transportation, and surveillance services has prompted the exploration of low-altitude airspace, that leads to Low-altitude economy (LAE) networks. Unlike traditional uncrewed aerial vehicle (UAV) systems, LAE envisions dense networks of flying platforms that serve both as mobile base stations and service nodes. However, the malicious deployment of UAVs in LAE networks can result in serious disasters. Therefore, robust and real-time UAV threat detection capabilities are required, particularly for low-signal-to-noise ratio (SNR) conditions. To address these challenges within LAE networks, we propose a Hybrid Quantum-Classical Convolutional Neural Network (HQC2) for low-SNR RF drone signal classification. The model fuses classical feature extraction with quantum variational circuits to leverage quantum superposition and entanglement for improved representation learning. By providing an efficient and noise-resilient RF sensing mechanism, the proposed HQC2NN directly supports the sensing plane of LAE architectures, enabling reliable situational awareness in dense, interference-prone environments. Simulations demonstrate a classification accuracy of 97.3%, outperforming classical counterparts under noisy conditions. The results underscore the potential of quantum-enhanced deep learning models for robust RF signal analysis and real-time drone detection.