AGRI-DRONE: Monitoring and Classification of Soil Fertility Based on Internet of Things Using Autonomous Drone

Indonesia is an agricultural country that faces significant challenges in maintaining soil fertility to support farming and plantation productivity, exacerbated by weather fluctuations and climate change. To address this issue, an Internet of Things (IoT) system called Agri-Drone was developed. The system is designed for soil fertility classification using fuzzy logic and weather prediction using machine learning, assisting farmers in making informed crop management decisions. Agri-Drone integrates components such as Soil Test, Weather Station, and LoRa Gateway carried by autonomous drones, as well as a website for monitoring, improving resource use efficiency, reducing the risk of crop failure, and supporting national food security. The system has demonstrated significant success, with the Soil Test and Weather Station components demonstrating Quality of Service (QoS) levels with end-to-end delay and response times of less than 10 seconds. The measurement accuracy of soil elements including nitrogen (N) was 91.93%, phosphorus (P) was 91.31%, potassium (K) was 88.7%, pH was 95.03%, and moisture was 93.54%, with Relative Standard Deviation (RSD) values showing high precision. The LoRa Gateway maintained a stable connection over a wide range, and the Weather Station showed a very high level of precision. Machine learning for weather classification achieved 98% accuracy, and weather prediction achieved 72%. The website is user-friendly with an average score of 4.584 on a scale of 5 with a total of 51 respondents, and achieved 100% performance according to GTMetrix, enabling effective monitoring of measurement results.