摘要
针对自动气象站采用的HMP45D型湿度传感器测量精度易受温度影响的问题,通过对遗传算法中的编码方式、适应度函数和参数进行改进研究,利用改进的遗传算法(genetic algorithm,GA)对反向传播(back propagation,BP)神经网络的初始权值阈值进行优化,在较大的范围进行搜索,采用反向传播算法在较小范围内进行微调,优化网络结构和参数,提出了用改进遗传算法优化BP神经网络的方法,根据在多温度条件下湿度传感器的实测数据,对利用此方法建立的模型进行温度补偿研究,并结合一般BP神经网络方法进行分析比较。实验结果表明,该方法具有全局寻优能力,补偿精度高,收敛速度快,能够有效补偿温度对湿度传感器的影响,大大提高了湿度传感器的测量准确度。
Aiming at the problem that the measurement accuracy of HMP45D humidity sensor adopted in automatic weather stations is vulnerable to the influence of temperature ,this paper studies and improves the encoding method,fitness function and parameters in genetic algorithm; uses the improved genetic algorithm （GA） to optimize the initial weights and threshold in back propagation（BP） neural network ;performs searching in larger range ;then uses the back propagation al- gorithm to carry out fine tuning in smaller range, and optimize the network and structure parameters at the same time. A new method is proposed,which uses the improved genetic algorithm to optimize BP neural network. Based on the measured data of humidity sensor in various temperature conditions, we carried out study on the temperature compensation model established using this method and performed analysis and comparison with general BP neural network method. The experimental results show that the proposed method has global optimization ability,high compensation precision,fast convergence speed;can effectively compensate the influence of temperature on humidity sensor and improve the measurement accuracy of humidity sensor greatly.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2013年第1期153-160,共8页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(61072133)
江苏省产学研联合创新资金计划(SBY201120033)
江苏省高校科研成果产业化推进项目(JHB2011-15)
江苏省"六大人才高峰"项目资助
关键词
遗传算法
BP神经网络
湿度传感器
GA—BP网络
温度补偿
genetic algorithm （GA）
back propagation（BP） neural network
humidity sensor
GA-BP neural net-work
temperature compensation
