Energy-Efficient Segment Clustering Algorithm for UAV trajectory

This paper proposes an energy-efficient clustering algorithm for UAV trajectory when a UAV scans massive IoT devices to collect data. The UAV trajectory power consumption in the scanning and data-collection process is mainly decided by the number of hovering points, path length, and collected data volume. These are then significantly affected by the number of grouped clusters of IoT devices and the amount of duplicate data collected from the repeatedly scanned IoT nodes in the overlapping areas of IoT clusters. Regarding this, we propose a low-complexity segment clustering (SC) algorithm aiming to appropriately group all the IoT devices into clusters with minimized overlap when considering the UAV communication range. The proposed SC algorithm is experimentally compared with existing clustering algorithms under five different topology scenarios. The numerical results show that the proposed SC algorithm outperforms its counterparts in most scenarios regarding the number of clusters, trajectory path length, and power consumption.