Flood Detection System Based On The Internet of Things (IoT) Using The Adaptive Neuro-Fuzzy Inference System (ANFIS) Method

Abstract

Floods are a natural disaster that often occurs in Indonesia, and have a major impact on material and non-material losses for the community. The absence of an early warning system and suboptimal handling exacerbates infrastructure damage and disrupts daily activities. To overcome these problems, this study designed a flood detection system based on the Internet of Things (IoT) integrated with the Adaptive Neuro-Fuzzy Inference System (ANFIS) method to improve accuracy in classifying potential flood risks. This system uses an HC-SR04 ultrasonic sensor to directly control the water level in the simulation container. This sensor is connected to the NodeMCU ESP8266 microcontroller to process and transmit water level data via the internet network. Data is displayed on the LCD screen and monitored in real time using the Blynk application on a smartphone. Data collection is carried out through simulation, where the higher the water (the closer to the sensor), the greater the risk of flooding. For example, when the distance of water to the sensor is 0 cm, the risk of flooding is 100%, and when the water level reaches 18 cm, the risk of flooding drops to 40%. The ANFIS method is used to classify flood status into three categories based on water level: Safe (18–11 cm), Alert (10–6 cm), and Danger (5–0 cm). The ANFIS model is drilled using simulation data to form membership functions and fuzzy rules. The test results show that the system is able to classify flood status accurately and provide early warning notifications to users. This system is expected to help communities and local governments respond to potential floods more quickly and accurately to reduce the impacts caused.