Forest Fire Detection and Guiding Animals to a Safe Area by Using Sensor Networks and Sound

Abstract

Forest fires are one of the main causes of environmental degradation. More than a million species of animals have lost their lives in the 2019–2020 wildfire that spread in the Amazon forest. The model that we are proposing, intends to drastically reduce the number of lives lost in such unfortunate events and also alert the first response accelerating their momentum. Our idea is to have Wireless Sensor Networks (WSN) placed in a widely distributed manner across the forest area. Each module consists of a smoke sensor, temperature, humidity sensor, and a speaker which is connected to a Node-MCU. These modules collect data that is necessary for the prediction of wildfires. The data collected is analyzed along with the wind direction by our deep learning algorithm which predicts the wildfire spreading direction. This prediction is used to find a safe route for the animals to move away and get to a safe zone. Then the animals are manipulated to move away from the wildfire with the help of distressing sounds produced from the speaker triggering their flight response for their self-preservation. These distressing sounds are produced in a pattern rather than just producing it wherever wildfire is present. Hence leading wildlife to a safe zone And also nearby villages can be warned by a siren.