An Intelligent Flood Automation System Using IoT and Machine Learning

Floods are one of the most common and well-known natural calamities that create havoc, causing damage to one's life and economy while adversely affecting the environment. This Research focuses on receiving emergency notifications whenever the situation is critical and also being able to predict the flood in time by automatically measuring the riverbed water level periodically. In this work, the objective is to measure the level using ultrasonic sensors and process the acquired data using ESP8266, where the information will be transferred to ThingSpeak IoT cloud which uses HTTP Protocol over Local Area Network thus enabling it to graphically monitor the world. It checks whether it is not exceeding the threshold limit and when it reaches beyond the limit the system should be able to alert individuals through Telegram and web cloud application, and when the system predicts a flood in the following timeline. Here the Servo Motor acts as a dam gate i.e., it opens when the flood is alerted and closes back once it reaches below the threshold value. For prediction, various machine-learning algorithms such as Random Forest, Support Vector Machine, KNN, Decision Tree, and Logistic Regression were implemented. Due to its high accuracy of about 95%, the logistic regression model was considered suitable for predicting the changes and detecting various possibilities.