Intelligent Smoke Detection: State Recognition and Monitoring of Heating Processes Using FMCW Radar and Data-Driven Algorithms

Conventional smoke detectors are prone to high false alarm rates, particularly in complex environments such as kitchens, where cooking smoke or water vapor often triggers alarms. This study explores the potential of frequency-modulated continuous wave (FMCW) radar technology combined with data-driven algorithms to enhance smoke detection accuracy and reliability. Unlike conventional methods, the proposed approach utilizes phase-based signal evaluation of fixed target measurements to monitor heating processes and differentiate between scenarios such as rising water vapor, lightly smoking oil, and fire. Experimental measurements at various heating levels were analyzed using an Extreme Gradient Boosting (XGB) classifier, achieving a test accuracy of 99.58% with a false alarm rate of 0%, demonstrating the capability to distinguish distinct heating processes based on their unique thermodynamic phase patterns.