Perpustakaan Digital ITB

The rapid growth of UAV (Unmanned Aerial Vehicle) research has enabled applications in infrastructure inspection. UAVs replace conventional methods that are time-consuming, costly, and dangerous. Equipped with cameras and sensors, UAVs inspect buildings in real-time, delivering high-resolution data for decision-making. Multi-drone coordination systems allow UAVs to collaborate through intelligent task distribution, increasing efficiency and coverage while minimizing overlap and energy consumption. Challenges remain in optimal path planning, dynamic task assignment, and coordination in complex building geometries. This research develops a multi-level optimization framework for autonomous drone swarm coordination in building inspection missions. The study addresses three key challenges: generating optimal flight paths that maximize coverage while minimizing distance, creating dynamic regional assignment strategies for balanced workload distribution, and designing adaptive waypoint generation systems for diverse building geometries. A comprehensive simulation environment was developed using PyBullet to evaluate algorithm performance across rectangular and octagonal building configurations with various swarm sizes. The proposed framework integrates enhanced path planning algorithms with Bayesian optimization for parameter tuning and dynamic point assignment strategies. Experimental results demonstrate significant performance improvements over baseline approaches, achieving coverage enhancements of 17.1%-63.1% in rectangular environments and 0.3%-46.5% in octagonal environments. Mission efficiency improvements up to 60.9% were observed in optimal configurations. The system successfully adapts to different building geometries while maintaining effective coordination through temporal-spatial overlap analysis and intelligent swarm sizing. These findings validate the effectiveness of the multi-level optimization approach for autonomous building inspection missions.