欢迎来到天天文库
浏览记录
ID:52207274
大小:391.95 KB
页数:6页
时间:2020-03-24
《基于PSO-BP神经网络的湿度传感器温度补偿.pdf》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、第28卷第6期传感技术学报V01.28No.6CHINESEJOURNALOFSENSORSANDACTUATORSJune20152015年6月TheTemperatureCompensationforHumiditySensorBasedonthePSo.BPNeuralNetworkXINGHongyan,,ZOUShuiping一,XUWei一,ZHANGQiang'(1.CollaborativeInnovationCenteronForecastandEvaluationofMeteorologicalDisasters,NangUniversityofIn
2、formationScienceandTechnology,Nang210044,China;2.KeyLaboratoryforAerosol—Cloud—PrecipitationofChinaMeteorologicalAdministration.NanjingUniversityofInformationScienceandTechnology,Nang210044,China)Abstract:AccordingtothetemperatureandhumiditysensorsofthetypeHMP45Dontheautomaticweatherstat
3、ionsinfluencedeasilybytemperatureintheactualapplication,compensationofhumiditysensormodelbytheBackProp-agation(BP)networkbasedonParticleSwarmOptimization(PSO)algorithmhasbeenproposed.TheinitialweightandthresholdofBPnetworkcanbesearchedgloballyinPSOalgorithm,thenassignstheoptimizedweighta
4、ndthresholdtoBPnetworkfortraining.Multiplegroupsofthehumiditysensordatashasbeenmeasuredunderthecon—ditionofdifferenttemperatures,Usingthismethodtoestablishamodelfortemperaturecompensation,andtheresultswerecomparedwithgeneralBPneuralnetworkmethod.Theexperimentalresultsshowthatthesumofabso
5、luteval—ueoferrorbytheuseofPSO-BPneuralnetworkmodelfortemperaturecompensationisreducedby10.3887%(RH)comparedwiththatofthetraditionalBPneuralnetworkmode1.PSO·BPneuralnetworknotonlycanovercomethelimitationsthatthetraditionalBPneuralnetworkiseasytofallintolocalminima,butalsohavethehigherpre
6、cision,anditcanmoreeffectivelycompensatetheinfluenceoftemperatureonhumiditysensor.Keywords:temperaturecompensation;particleswarmoptimizationalgorithm(PSO);BPneuralnetwork;humiditysensorEEACC:7230doi:10.3969/j.issn.1004—1699.2015.06.015基于PSO—BP神经网络的湿度传感器温度补偿冰行鸿彦',邹水平,徐伟。,张强(1.南京信息工程大学气象灾害
7、预报预警与评估协同创新中心,南京210044;2.南京信息工程大学,江苏省气象探测与信息处理重点实验室,南京210044)摘要:针对自动气象站采用的HMP45D型温湿一体化传感器在实际应用过程中易受温度影响的问题,提出了基于粒子群优化算法(PSO)的BP神经网络温度补偿模型,利用粒子群优化算法对BP神经网络的初始权值阈值进行全局寻优,将粒子群优化算法优化好的权值阈值赋给BP神经网络,对BP神经网络进行训练。根据不同温度条件下测得的多组湿度传感器数据,通过建立模型,实现温度补偿,与传统BP神经网络补偿结果进行比较。实验表明,与传统BP神经网
此文档下载收益归作者所有