DESIGN AND EXPERIMENTAL EVALUATION OF A DUAL AGENT FIREFIGHTING ROBOT WITH AUTOMATIC FIRE CLASS SELECTION FOR HAZARDOUS INDOOR ENVIRONMENTS

Abstract

Fire accidents present significant dangers to people’s lives and physical assets due to unsafe and high-risk conditions which constrain risk-free and efficient fire suppression efforts. Thus, an intelligent robotic firefighting system is required to operate in high-risk environments without compromising human operators’ safety. This paper describes the design, fabrication, and experimental testing of a low-cost dual agent firefighting robot with automatic fire class selection by rule-based sensor fusion. The robot identifies fire, categorizes it and picks a suitable extinguishing agent without the operator intervention. An Arduino Nano is coupled with an infrared flame sensor, a smoke sensor and a camera to carry out real time fire detection and classification. Water based pump is used to suppress Class A fires and CO₂ or foam are used to suppress Non-Class A fires. Controlled indoor experimental results indicated a classification accuracy of 100 percent with Class A fire and 95 percent with Non-Class A fire. Class A fires extinguished in an average of 41 seconds and Non-Class A fires extinguished in an average of 12 seconds with a reduction of flame area of more than 80 percent. The results demonstrate the feasibility of an affordable dual agent robotic system that can be used in confined indoor firefighting.