<![CDATA[Fires are disasters that can occur suddenly and cause significant losses to both human lives and property. Conventional detection systems are generally limited to smoke or temperature alarms, making them less accurate and often slow in providing warnings. This study aims to design and implement an Internet of Things (IoT)-based fire detection system capable of providing real-time early fire warnings. The system employs an ESP32 microcontroller connected to a DHT22 sensor to measure temperature and humidity, an MQ-2 sensor for flammable gas detection, and a flame sensor for direct flame detection. Sensor data are transmitted to the Firebase Real-time Database and forwarded as automatic notifications via the Telegram application. Testing was conducted on 400 data samples covering both fire and non-fire conditions, with performance evaluation using Precision, Recall, Accuracy metrics, and End-to-End Delay measurement. The results show an accuracy of 91%, precision of 88.55%, recall of 93.19%, and an average notification delivery delay of 2.804 seconds. These findings indicate that the system can detect fires effectively and responsively on a small scale, with potential for further development through additional sensors and testing in more complex environments.]]>
Copyrights © 2025
