T-S型RBF神经网络在红外火焰探测系统中的应用

基于T-S(Takagi-Sugeno,高木-关野)模型RBF(Radial Basis Function,径向基函数)神经网络,提出了一种应用于三波段点型红外火焰探测器的识别算法,同时实现了硬件电路以及软件程序的设计。针对火焰探测器在检测过程中可能出现的数据丢失、失真、饱和等复杂情况,本文利用RBF网络较优的逼近精度和泛化能力,同时结合T-S模型用少量的模糊规则可生成较复杂的非线性函数的特点,实现了火焰与干扰源的准确识别。实验证实,T-S模型RBF神经网络相比于BP(Back Propagation,反向传播)网络在逼近精度、收敛速度、鲁棒性等多个方面都有所提升。

Application of T-S RBF neural network in infrared flame detection system

In this paper,based on the Takagi-Sugeno(T-S)model Radial Basis Function(RBF)neural network,a recognition algorithm for three-band point-type infrared flame detector is proposed,and the hardware circuit and software program are designed. In view of the data loss,distortion,saturation and other complex situations that may occur in the process of flame detector detection,accurate identification of flame and interference source is realized in this paper utilizing the better approximation accuracy and generalization ability of RBF network,and the characteristics of T-S model,which can generate more complex nonlinear functions with a few fuzzy rules. Experiments show that using T-S model RBF neural network is better than the BP neural network in approximation accuracy,convergence speed,robustness and other aspects.