Fusion of Ship Fire Information Based on Fuzzy DS Evidence Theory

A Dempster-Shafer evidence theory multi-source information fusion technology is proposed to address the issues of missed and false alarms in early prediction of ship fires, which integrates fuzzy and unified trust allocation formulas as the framework for calculating Basic Probability Assignment. Using PyroSim to establish an indoor model of ships for fire simulation, obtaining the trust of sensors through Sigmf function and unified allocation formula. After multi-sensor and multi period fusion, the probability of fire occurrence is determined to be 0.354 higher than the highest probability of a single sensor initially. This improves the accuracy and effectiveness of fire identification, improves the accuracy of fire alarm, and is of great significance for achieving intelligent fire alarm in ships.