Path Planning Algorithms for UAV Precision Agriculture: Methods, Gaps, and Future Directions

Abstract

Unmanned Aerial Vehicles (UAVs) are increasingly central to precision agriculture, enabling high-resolution monitoring and targeted interventions across heterogeneous fields. Effective deployment hinges on path-planning algorithms that deliver safe, energy-aware trajectories under real farm constraints such as wind, irregular canopies, no-fly zones, moving machinery, and tight battery budgets. This paper surveys and synthesizes path-planning approaches for agricultural UAVs across four families: classical graph search such as heuristic and sampling, machine learning based, deep learning and others. The study maps each family to agricultural objectives such as coverage efficiency, safety, smoothness, constraints (endurance, turning limits, communication), and deployment readiness. The review identifies persistent gaps: heavy reliance on simulation, limited field validation in dynamic environments, weak coupling between planners and energy/endurance models, and scaling challenges for multi-UAV coordination. The study also highlights converging trends, hybrid DRL with sampling planners, heuristic coverage with agronomic constraints, perception-driven planning using lightweight vision models, and emerging use of Dynamic Mode Decomposition (DMD) for model reduction and onboard control. Finally, the study outlines a research agenda toward adaptive, field-validated planner that integrate energy modeling, multi-robot coordination, and onboard-deployable perception, providing a roadmap for robust autonomy in precision agriculture.