Algoritma Pengendalian Kelembapan Tanah Menggunakan MQTT dan Fuzzy Logic untuk Peningkatan Produktivitas Pertanian

Abstract

To optimize soil moisture management in agricultural settings, this study introduces a smart irrigation system that integrates MQTT communication with a fuzzy logic controller. The system’s scalable and flexible design allows for the straightforward addition or removal of sensors and actuators without modifying the core architecture. Key environmental variables—soil moisture, air temperature, and air humidity—drive the decision-making process, which is handled by a fuzzy inference system refined through decision tree pruning. This approach maintains inference accuracy while simplifying the rule set. The lightweight publishsubscribe model provided by MQTT supports rapid, real-time data exchange. Additionally, MQTT’s hierarchical topic structure and single-level wildcards (+) enable efficient management of multiple zones and devices, making integration and expansion seamless. Simulation results demonstrate low communication latency and consistent system performance, contributing to improved water efficiency and higher crop yields. The system also achieved a match rate of over 97% with manual decisions, confirming the logical soundness of its recommendations. Network analysis using Wireshark recorded an average MQTT message round-trip time of 36.76 ms, well below the ITU-T G.114 real-time threshold of 150 ms.