Optimizing UAV sprayer performance using field data and machine learning approaches
Unmanned Aerial Vehicles (UAVs) have become increasingly relevant in precision agriculture, particularly for plant protection applications. However, optimizing UAV spray performance remains a challenge due to the complex interactions between droplet size, flight parameters, and environmental conditi...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-08-01
|
| Series: | Smart Agricultural Technology |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2772375525002461 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Unmanned Aerial Vehicles (UAVs) have become increasingly relevant in precision agriculture, particularly for plant protection applications. However, optimizing UAV spray performance remains a challenge due to the complex interactions between droplet size, flight parameters, and environmental conditions. This study aims to enhance spray deposition efficiency by integrating field experiments, statistical evaluations, and based on physics simulation models. Experimental data were collected under actual conditions field conditions using four droplet size categories (60, 100, 140, and 180 μm) and two flight altitudes (2 m and 3 m) at speeds of 5.5 m/s and 6.9 m/s. Coverage rate, real droplet size, and deposition uniformity were assessed. Statistical tests (Levene’s test, Kruskal-Wallis, and Dunn’s test) identified significant differences among experimental groups. Machine learning models (Random Forest and XGBoost) were employed to evaluate parameter importance and predict real droplet size and coverage performance. Additionally, particle-based simulations were conducted in Python to visualize droplet deposition patterns under six UAV flight scenarios. Results showed that flight speed and droplet size were the most influential factors affecting spray coverage. Among the tested configurations, the 2 m / 5.5 m/s combination provided the best balance between coverage rate and distribution uniformity. The heatmaps based on simulation confirmed and visually supported the experimental findings. This integrated approach demonstrates the potential of combining field data with simulation and machine learning to optimize UAV spraying strategies, offering a framework for broader applications in sustainable agriculture. |
|---|---|
| ISSN: | 2772-3755 |