Research Of Intelligent Fire Alarm System Based On Fuzzy Neural Network

Fire is a common disaster, which could cause significant threats and losses to human life,and property. The world has been working on the internal mechanism of fire study designed to prevent fires and minimize losses caused by fire. Among the many means, the fire detection technology can effectively prevent property and reduce the loss. Traditional fire detection technology is based on a single parameter, fault alarm and leakage alarm is very common, this paper focused on research based on fuzzy control and neural network intelligent fire detection systems, fire detection algorithm has been made some progress, main content as follows:(1)the traditional fire detection technologies have many disadvantages which can not effectively meet the needs of today's fire alarm system, a fuzzy reasoning and neural network intelligent fusion has been proposed,using CO content, environmental temperature, smoke content for the system input,deploying artificial intelligence technology for multi-signal integration;(2) using fuzzy logic characteristics of strong capabilities dealing nonlinear problems, through fuzzy,fuzzy reasoning and anti-fuzzy to get 64 fuzzy rules, the algorithm simulation using MATLAB implementation of fuzzy reasoning as an important neural network output secondary criteria;(3)BP and RBF networks were used to establish neural network model of fire detector, designed with the MATLAB language in the fire detection system algorithm for neural networks, comlete network of training, Simulation results were comared;(4) organically combine Emirical knowledge-based fuzzy logic and a high degree of flexibility, self-learning ability of neural networks, to maximize their own characteristics and effect on the outut of decision-making resents a new criteria, greatly imrove the accuracy of fire detection;(5) using MATLAB M-files, and VC++mixed rogramming, intelligent fusion algorithm and the organic integration of the visual interface, the system will eventually be making the results dislayed in human machine interface intuitive;(6) The availability and reliability of intelligent fire alarm system algorithm simulation results manifest that the system can distinguish between non-fire source of smoldering fires and fire with high precision, and system has high sensitivity, reducding errors and false alarm rate efficiently. The system has high accuracy and high reliability, while it still need to be further imroved on fire sample database.