UAV-Aided 3D Spectrum Mapping Framework: Architecture, Technique, and Opportunity

Spectrum mapping (SM) visualizes spectrum information across geographical regions through radio environment maps (REMs), enabling critical applications in spectrum monitoring, management, and security. In dynamic space-air-ground integrated networks, unmanned aerial vehicles (UAVs) can extend the sensing coverage of three-dimensional (3D) SM beyond traditional ground systems. However, constructing accurate REMs in dynamic and unknown 3D environments remains challenging due to sparse, noisy measurements, and nonstationary propagation conditions. This article proposes a comprehensive UAV-aided 3D SM framework, integrating architecture, key techniques, and practical opportunities. It features: 1) Adaptive air-to-ground channel modeling using a data-driven channel dictionary to capture realistic propagation dynamics; 2) Intelligent UAV path planning combining realtime sensing with an information-driven decision framework for efficient data acquisition; 3) Robust spatial completion via sparse Bayesian learning to ensure accurate recovery of unsensed regions; and 4) Channel model-guided deep learning for temporal prediction of spectrum evolution. A prototype system is implemented and validated in both simulations and field measurements. Applications in dynamic spectrum access, interference management, and UAV navigation are discussed, along with open challenges in data fusion, real-time processing, and adaptive modeling. This work provides a foundation for the 6th-generation (6G)-enabled spectrum management and UAV-assisted wireless networks, bridging theory and practical deployment.