基于微粒群算法和模糊神经网络的火灾探测研究

针对火灾探测的特点,将模糊系统和神经网络有机结合,实现模糊系统设计参数的自动调整。采用符合国家标准明火、阴燃火以及厨房环境下的干扰火等作为模糊神经网络的训练样本和测试样本,依据模糊神经网络算法要求,完成了网络结构的设计,并给出相应的计算模型,利用微粒群算法对网络的权值进行学习与训练。结果表明,该算法在探测国家标准火的火灾状态方面具有有效性和可行性。

Fire detection based on fuzzy neural network combined with particle swarm optimization algorithm

An intelligent algorithm for processing fire detection signal is proposed aiming at the characteristic of fire detection.The automatic adjusting of design parameter in fuzzy system is realized by integrating fuzzy system and neural network.According to the request of fuzzy neural network algorithm,the design of network configuration is accomplished and calculating model correspondingly is given,then use particle swarm optimization algorithm completes the network's weights' learning and training.The result shows the algorithm is validity and feasibility.