摘要
针对室内多点火灾探测预警需求,设计新型火灾预警系统,从机处理器DSP28335通过传感器矩阵采集6种物理量,进行单BP神经网络预警识别,多个从机由CAN总线组网,将数据传输到主机,存储的同时通过串口、以太网与PC上位机Matlab通信,完成仿生态神经网络群BNNG聚类预警识别,将结果反向传输给从机和主机处理器,执行具体动作,经过测试,系统正确识别率可以到达95%,达到实际应用需求。
A new type of fire warning system is designed for aiming at the demand of multi point fire detection and prewarning, 6 kinds of physical quantities are collected through the sensor matrix by the slave computer＇s processor DSP28335, and the single BP neural network early-warning recognition is carried out.Multiple slave computers are connected by CAN bus,the data is transmitted to the host computer by the serial port or the Ethernet to the Matlab software which completes the prewarning recognition of the Bionics Neural Network Group, and transmits the results back to the slave computer and the host processor to perform specific actions. After testing, the correct recognition rate of the system can reach 95%, and the actual application needs are reached.
作者
沈晓波
王留留
蔡俊
SHEN Xiaobo;WANG Liuliu;CAI Jun
出处
《淮南师范学院学报》
2018年第2期126-129,共4页
Journal of Huainan Normal University
基金
安徽省自然科学重点项目"室内火灾分级预警BNNG算法关键技术研究"(KJ2017A455)
校重点项目"通用型组网式火灾概率数据分析预警仪研制"(2016xj01zd)
关键词
分级预警
BNNG算法
归一化
聚类
classification prewarning
BNNG algorithm
normal ization
clustering
