论文部分内容阅读
提出将ＡＲＴ－２神经网络与模糊系统串联而成的模糊ＡＲＴ－２神经网络系统并用于火灾探测，使其对火灾信号具有很好的聚类特性．实验结果表明，该系统比ＢＰ神经网络火灾探测系统具有更强的适应环境的能力，能够更快、更准确地探测各种标准试验火并具有很强的抗干扰能力．
This paper proposes a fuzzy ART-2 neural network system, which is composed of ART-2 neural network and fuzzy system, and is used in fire detection to make it have good clustering characteristics for fire signals. Experimental results show that the proposed system has a better ability of adapting to the environment than the BP neural network fire detection system, and can quickly and accurately detect all kinds of standard test fires and has strong anti-interference ability.