Self-learning fuzzy neural network and its application to fire auto-detecting in fire protection systems

A fuzzy-neuro network architecture is proposed. An initial fuzzy knowledge base can be built by means of self-adaptive learning from historical data, fuzzy rules and system parameters can be optimized by online learning of the FNN via real-time data, which make the system possess distinguished adaptive features and self-learning capability. Taking fire auto-detecting in a fire protection system as application background, a series of experimental research has been carried out. Experimental results demonstrate the feasibility of the neuro-fuzzy system.