Evidence-Based Artificial Intelligence and Internet of Things Systems for Smart Fire Detection and Firefighting in Oil Companies: A Comprehensive Review

Abstract

Fire accidents are considered one of the major threats to the survival and safety of humans and the surrounding infrastructure and environment, especially
in industrial and smart building scenarios, and large outdoor spaces such as oil and gas industries. The traditional fire detection and firefighting systems, which are
mostly based on sensors such as smoke, heat, and flames, are often reported to have poor response time, awareness level, and high rates of false alarms. However,
with the recent developments in Artificial Intelligence (AI) and Internet of Things (IoT) technology, intelligent systems have now become feasible for fire detection
and firefighting. In this paper, an exhaustive systematic review is proposed to analyze and study AI-based intelligent systems for fire detection and firefighting by
categorizing them according to their methodologies, devices used, environments, advantages, and limitations. A taxonomy is proposed to categorize the existing
AI-based systems into vision-based systems, sensor-based systems, IoT-based systems, edge AI-based systems, UAV-based systems, robotic systems, and hybrid
multi-modal systems. The paper also discusses the recent developments in AI-based deep learning-based fire detection systems and intelligent sensors and IoTbased systems. Despite the progress made in the field of technology, there are some issues that need to be addressed, such as the availability of standard fire
datasets, computational costs, environmental variations, cybersecurity threats, and the absence of large-scale real-world deployments. Moreover, the current
research indicates that there is an absence of multi-modal sensing, explainable AI, and edge computing for the development of real-world firefighting applications.
The gaps identified in the research are essential for the development of intelligent fire detection and firefighting applications, including multi-modal AI, gasistorbased intelligent sensing, explainable fire detection, and edge computing for firefighting applications. The findings of the study are essential for the development
of intelligent fire detection and firefighting applications.