My student Estela Mora Barba defended her Electronic Communications BSc thesis on embedded AI for precision agriculture
🎓 Outstanding Bachelor’s Thesis Defense
On Friday, September 19, 2025, my student Estela Mora Barba successfully defended her Bachelor’s Thesis entitled “Device for Generating Prescription Maps Using Embedded AI in Precision Agriculture” at the ETSIST–Universidad Politécnica de Madrid. Her project was supervised within the GRyS Research Group, building upon the European initiatives AFarCloud and FlexiGroBots.
The work presents a fully autonomous UAV-based embedded system capable of capturing agricultural imagery, processing it locally using AI inference, and generating prescription maps in real time, without depending on external cloud infrastructure. This approach advances the autonomy, efficiency, and sustainability of precision agriculture operations through edge computing and modular microservice architecture.
From a technological standpoint, the device integrates RGB cameras, an embedded computing module with Hailo-8 and Hailo-8L AI accelerators, and ROS 2 Jazzy Jalisco as middleware for on-board mission control. The architecture employs gRPC for inter-service communication, REST APIs for external interfaces, Docker containers for deployment portability, and ThingsBoard for telemetry and visual analytics.
The system connects seamlessly with the FleetManager platform, enabling dynamic coordination of multiple UAVs and efficient mission scheduling across agricultural environments. Validation was carried out following ISO/IEC/IEEE 29119:2020, covering functional, performance, and reliability testing. Results confirmed the system’s robustness and suitability for real agricultural missions.
Economically, the solution relies on low-cost hardware and open-source software, ensuring affordability and replicability. Environmentally, it enhances sustainable resource management by enabling precise treatment application and early disease detection—directly contributing to several UN Sustainable Development Goals.
Future extensions include integration with autonomous ground vehicles (UGVs), improved AI models for multi-crop support, and multispectral analysis for finer agronomic diagnostics. This project exemplifies the potential of embedded AI and robotics to transform agri-tech toward smarter, greener, and more autonomous farming.