Date of Publication :15th March 2025

Abstract: Fire accidents pose a significant threat to human life and property, making the development of predictive fire detection systems essential. Traditional systems rely primarily on gas sensors, detecting fires only after significant spread, which can result in severe property damage or even life-threatening situations. Our AI-driven fire detection system aims to address this by using a combination of sensors, including color, temperature, and gas sensors, to detect fires at an early stage. A neural network is implemented on the Arduino Nano 33 BLE Sense, enabling the microcontroller to predict fires based on proximity and color intensity patterns. Alerts are sent to users via a mobile app, while relevant authorities are notified through messaging protocols. The system also integrates a central server for data analysis and continuous improvement, enhancing prediction accuracy over time . This proactive approach ensures faster response times, minimizing damage and improving safety in various environments.

Reference :