Article Preview
Top1. Introduction
Every year, there are countless casualties and property losses caused by fire disasters in various countries. With the advent of the information age, society has created more material wealth, and people's quality of life has improved. However, the occurrence of fires is also on the rise year by year, and the risk coefficient is increasing. Fires have brought great disasters to humanity (Sowah & Ofoli et al., 2017). The fire not only brought tremendous pressure to the victims economically, but also caused severe trauma to the victims psychologically (Biase & Laneve, 2018). People put higher requirements for the working and living environment, and this objective social demand has promoted the growth of traditional buildings to intelligent buildings (Silvani & Morandini et al., 2015). This kind of large-scale intelligent building with concentrated personnel and wealth will have disastrous consequences in case of fire, so higher requirements are put forward for the fire fighting system. The traditional fire alarm and fire fighting linkage system has revealed many shortcomings in high-rise and intelligent buildings (Cakiades & Deligeorges et al., 2015). Due to the rapid growth of economic construction, the continuous improvement of people's living standards, and the increasing shortage of urban land, it has promoted the continuous improvement of urbanization, the sharp increase in crowded places, and the increasing probability of fire. Realizing timely detection and alarm of fires has become a very concerned issue (Dash & Tirtharaj, 2017).
The automatic fire alarm system is an indispensable safety facility for modern fire fighting. It is an essential guarantee for people to find fires early and take effective measures to control and extinguish them in time (Hisham & Hamdy et al., 2021). The most commonly used fire monitoring sensors are based on fire characteristics such as heat, smoke and gas. Because each type of fire sensor has a specific application occasion and scope, and the fire monitoring environment is complex and diverse in practical application, the false alarm and missing alarm rate of automatic fire alarm system with a single type of sensor is very high (Nemalidinne & Gupta, 2018). The traditional fire alarm control system is developing in the direction of high intelligence. Its typical characteristics are high reliability, low false alarm rate and a high degree of automation. Data fusion is the intelligent synthesis of multi-source information from the system, resulting in more accurate and complete estimation and judgment than a single information source (Nguyen & Park et al., 2018). It can not only overcome the instability and limitation of a single sensor, but also grasp the system information more comprehensively and accurately (Wang & Du et al., 2020). ANN has large-scale parallelism, redundancy, fault tolerance, nonlinear mapping and high self-learning, self-organization and adaptive ability (Dora & Subramanian et al., 2016). In this paper, a multi-sensor data fusion algorithm based on ANN technology is proposed, which intelligently processes various environmental characteristic parameters detected by multi-sensors, effectively detects accurate fire signals, and realizes early fire monitoring and alarm.