<![CDATA[Flood disasters remain a recurring hydrometeorological problem in Indonesia, requiring adaptive and reliable early warning systems. This study proposes an Internet of Things (IoT)-based Flood Early Warning System (FEWS) utilizing the Mamdani Fuzzy Inference System to classify flood risk levels. The system integrates three hydrological input parameters: rainfall intensity, water level, and flow rate, acquired through real-time sensors connected to an Arduino Uno microcontroller. A total of 36 fuzzy inference rules were constructed from the permutation of three rainfall sets, three water level sets, and four flow rate sets. The defuzzification process applies the Centroid method to generate a crisp flood alert value. Experimental testing using rainfall of 48 mm, flow rate of 40 m³/s, and water level of 145 cm produced a defuzzification value of 131 (MATLAB simulation) and 130.8 (hardware prototype), both classified under the “Normal” category. The minor numerical deviation (0.2%) confirms system consistency across platforms. The results demonstrate that integrating multi-parameter hydrological variables enhances decision resolution compared to conventional threshold-based systems. Therefore, the proposed model provides an adaptive and reliable approach for real-time flood risk mitigation.]]>
Copyrights © 2026
