AI-Enabled Swarm Clustering for UAVs-Aided IoT Data Acquisition

This paper presents a novel artificial intelligence (AI)-enabled framework for optimizing data acquisition in Unmanned Aerial Vehicle (UAV)-aided Internet of Things (IoT) networks through intelligent swarm clustering. The proposed framework addresses the challenge of efficient data collection from distributed IoT devices (IoTDs) by employing a two-stage optimization approach: a K-Means-based clustering algorithm for IoTD association and a data-weighted centroid optimization strategy for UAV positioning. We introduce a dynamic clustering mechanism that considers spatial distribution and data volume heterogeneity of IoTDs to determine optimal UAV hovering locations. Extensive simulations demonstrate the superior per-formance of our proposed framework compared to conventional approaches. Results show that our method achieves up to 41.5% reduction in data collection time compared to random clustering with greedy UAV positioning (RC-GU) and maintains consistent performance advantages across various transmission power levels. The framework exhibits particular efficiency in low-power scenarios, making it especially suitable for energy-constrained IoT environments. This work advances intelligent UAV swarm management and provides a promising solution for large-scale IoT data acquisition scenarios.