Optimizing Agricultural Irrigation with a Three-Layer Neural Network–Driven Adaptive Drone Sprayer

Autonomous drone technology continues to advance, with growing implementation in the agricultural sector to enhance the efficiency of irrigation and crop maintenance. One notable innovation in this field is the adaptive spraying system, which leverages a Tri-layered Neural Network to dynamically adjust water discharge based on real-time environmental conditions such as wind speed, flight altitude, and soil moisture. While this technology has demonstrated significant improvements in water-use efficiency, its performance under highly dynamic environmental variations still requires further evaluation. Preliminary findings indicate that wind speed and flight altitude exert the most substantial influence on irrigation accuracy and precision, while GPS error also plays a critical role in the drone’s ability to accurately target designated areas. This study aims to comprehensively evaluate the performance of an autonomous drone equipped with a machine learning–based adaptive sprayer under varying environmental conditions. The dataset analyzed includes parameters such as wind speed, drone speed, flight altitude, GPS error, wind gust intensity, and soil moisture. Experimental results reveal that the Tri-layered Neural Network can classify irrigation output into all predefined levels with an accuracy exceeding 95%, underscoring its effectiveness and strong potential for optimizing agricultural irrigation practices.