Integration of Fuzzy Logic and Machine Learning for Real-Time Fire Detection: A Systematic Review

Resumo

This systematic review critically examines the integration of fuzzy logic and machine learning algorithms as a strategy to enhance precision in fire detection and optimize real-time response, particularly in environments characterized by high complexity and operational variability. A total of 50 studies from diverse geographic and application domains were analyzed, focusing on hybrid architectures involving multisensor fusion, neuro-fuzzy inference systems, and integration withInternet of Things (IoT)-enabled technologies. The comparative analysis encompassed hybrid algorithms, sensor fusion strategies, adaptive response mechanisms, and the incorporation of environmental variables in modeling detection systems. 
Findings indicate that the convergence of fuzzy logic and machine learning can elevate system accuracy to above 90%, substantially reduce false alarm rates, and dynamically adapt to various operational scenarios through self-learning models. Integration of temperature, smoke, carbon monoxide, and visual sensors enhanced reliability, while deployment in embedded IoT platforms reduced detection latency by as much as 70%, enabling more timely and effective responses.Despite these promising advancements, most validations rely on controlled simulations, with limited real-world implementations. The heterogeneity of fuzzy methodologies further challenges standardization and result reproducibility. Future studies should prioritize field validations and develop standardized frameworks to support scalability, interoperability, and economic viability across different application contexts.