A Hybrid Framework for Fire Outbreak Detection Based on Interval Type-2 Fuzzy Logic and Flower Pollination Algorithm

Abstract

In this paper, a hybrid approach for fire outbreak detection based on interval type-2 fuzzy logic (IT2FL) and flower pollination algorithm (FPA), using environmental parameters is proposed. Due to the high uncertainty in the fire outbreak data, IT2FLSs are able to consider many linguistic uncertainties in the membership functions (MFs) of the type-2 framework, thus, it can raise the accuracy of the fuzzy system. The MF parameters of IT2FL controller are optimized by the flower pollination algorithm (FPA). The fire outbreak data capturing device (FODCD) is developed to extract fire outbreak environmental data and store in a database. The proposed optimized controller is simulated and experimentally applied to detect a fire outbreak. The controller performance is compared with the conventional type-2 fuzzy logic-based controller, respectively, in the MATLAB/Simulink environment. The experimental result indicates that with the temperature at 40.657 °C, smoke at 77.86%, flame at 762.95 ppm (part per million) and T (threshold) of 0.8, the IT2FL-FPA and IT2FL predict fire outbreak with 0.8276642 (83%) and 0.777972 (78%) possibility, respectively. The performance results show that when the threshold T is kept between an optimal range of 0.8 and 0.85, the IT2FL-FPA model gives accuracy between 93.33% and 100% with an error rate between 0 and 0.07%, while the IT2FL model gives accuracy between 90 and 96.67% with an error rate between 0.03 and 0.1%. The simulation and experimental results show that the IT2FL-PFA controller outperforms the same controller without optimization.