Abstract:
This thesis presents the development of an autonomous quadcopter designed specifically for date-
palm pollination. The work integrates a complete nonlinear dynamic model, a cascaded control
architecture for stable low-speed flight, and a lightweight autonomous navigation strategy adapted
to irregular oasis environments. A YOLOv8-based perception system enables real-time detection
of palm flowers, while a custom hardware platform—combining an ESP32 flight controller,
Raspberry Pi Zero 2 W, and a pollen pump—forms the operational prototype. Simulation and
preliminary experimental results show that the proposed system can achieve precise hovering,
reliable flower detection, and safe maneuvering near palm crowns. The findings demonstrate the
technical feasibility of UAV-based pollination and highlight its potential to reduce labor demands,
improve safety, and enhance consistency in date-palm agriculture.
