Path planning of agricultural UAVs for combined coverage and spot spraying application

Abstract

This thesis presents the development of a path planning algorithm for unmanned aerial vehicles (UAVs) to enhance the precision and efficiency of agricultural weed management through combined spot and blanket spraying. Motivated by the growing demand for sustainable and resource-efficient farming, the proposed approach leverages unsupervised clustering, spatial statistics, and optimization techniques to identify weed-dense regions and minimize non-target spraying. Incorporating Density-Based Spatial Clustering of Applications with Noise (DBSCAN) with principal component analysis (PCA) and adapting the traveling salesman problem (TSP) algorithms, the system generates optimized UAV trajectories that reduce operational time, chemical usage, and ecological impact. The algorithm demonstrates that the integrated path planning methodology achieves less than 1.5 times the optimal route. Simulation and field datasets show a significant reduction in non-target spraying while maintaining weed knockdown efficacy better than conventional blanket spray approaches. Additionally, the framework addresses practical UAV constraints, including coverage limits, and adaptability for heterogeneous field conditions. This research presents a scalable solution for autonomous agricultural spraying, enabling data-driven farm management, promoting sustainable crop production, and aligning with global trends in precision agriculture and agricultural automation.