The IoT-based indoor fire extinguishing system with the C4.5 algorithm was developed to improve early detection, real-time monitoring, and automatic extinguishing at room scale. The system uses an ESP32, MQ-135, flame infrared, DHT22, and ultrasonic sensors to read environmental conditions, then sends the data to Firebase Realtime Database and an Android application. The C4.5 model uses a five-point time window to predict Safe, Alert, and Danger states. Training results show an average accuracy of 84.9% and best performance above 90%. Functional testing on 20 scenarios produced 13 successful and 7 failed cases, with a 65% success rate. Failures were mainly caused by mechanical belt-conveyor constraints, indicating that stronger mechanical components should be prioritized in future development.
Copyrights © 2026