CN117686555B - A drift compensation method for LC humidity sensor based on machine learning - Google Patents

A drift compensation method for LC humidity sensor based on machine learning Download PDF

Info

Publication number
CN117686555B
CN117686555B CN202410153167.5A CN202410153167A CN117686555B CN 117686555 B CN117686555 B CN 117686555B CN 202410153167 A CN202410153167 A CN 202410153167A CN 117686555 B CN117686555 B CN 117686555B
Authority
CN
China
Prior art keywords
data
model
humidity sensor
neural network
function
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202410153167.5A
Other languages
Chinese (zh)
Other versions
CN117686555A (en
Inventor
任青颖
魏鸿飞
郭宇锋
姚佳飞
李金泽
李卫
许杰
王德波
许巍
王萌
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University Of Posts And Telecommunications Nantong Institute Co ltd
Nanjing University of Posts and Telecommunications
Original Assignee
Nanjing University Of Posts And Telecommunications Nantong Institute Co ltd
Nanjing University of Posts and Telecommunications
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing University Of Posts And Telecommunications Nantong Institute Co ltd, Nanjing University of Posts and Telecommunications filed Critical Nanjing University Of Posts And Telecommunications Nantong Institute Co ltd
Priority to CN202410153167.5A priority Critical patent/CN117686555B/en
Publication of CN117686555A publication Critical patent/CN117686555A/en
Application granted granted Critical
Publication of CN117686555B publication Critical patent/CN117686555B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N27/00Investigating or analysing materials by the use of electric, electrochemical, or magnetic means
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/10Pre-processing; Data cleansing
    • G06F18/15Statistical pre-processing, e.g. techniques for normalisation or restoring missing data
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/21Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2131Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on a transform domain processing, e.g. wavelet transform
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/0499Feedforward networks
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods
    • G06N3/084Backpropagation, e.g. using gradient descent
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods
    • G06N3/086Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02Preprocessing
    • G06F2218/04Denoising
    • G06F2218/06Denoising by applying a scale-space analysis, e.g. using wavelet analysis
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08Feature extraction
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02TCLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00Road transport of goods or passengers
    • Y02T10/10Internal combustion engine [ICE] based vehicles
    • Y02T10/40Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Chemical & Material Sciences (AREA)
  • Physiology (AREA)
  • Probability & Statistics with Applications (AREA)
  • Chemical Kinetics & Catalysis (AREA)
  • Electrochemistry (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Investigating Or Analyzing Materials By The Use Of Electric Means (AREA)

Abstract

The invention discloses a machine learning-based LC humidity sensor drift compensation method, which comprises the following steps: carrying out a temperature test experiment on the LC humidity sensor to obtain original data, wherein the original data comprises a plurality of groups of data, and each group of data comprises response data under humidity and temperature conditions; preprocessing response data, extracting response bandwidth BW, real part impedance maximum Re #) Max and the resonance frequency Freq corresponding to the maximum value thereof, denoising the Max by utilizing wavelet analysis, and carrying out normalization processing on the denoised data to be used as a characteristic to establish a data set; initializing a BP neural network, and optimizing the BP neural network model by adopting a genetic algorithm to obtain a GA-BP model; the GA-BP model is trained with the training set, and the trained model is used for compensation of the drift data test set. The invention compensates the drift of the LC humidity sensor through the neural network model, and the model is obtained through learning the drift data setThe coefficient reaches 0.966, the model can explain 96.6% of uncertainty, and good compensation effect is obtained.

Description

一种基于机器学习的LC湿度传感器漂移补偿方法A drift compensation method for LC humidity sensor based on machine learning

技术领域Technical Field

本发明涉及LC传感器技术领域,具体涉及一种基于机器学习的LC湿度传感器漂移补偿方法。The present invention relates to the technical field of LC sensors, and in particular to a drift compensation method for an LC humidity sensor based on machine learning.

背景技术Background technique

LC无源无线传感器本质上是一种LC谐振环路,由无源元件电阻、电容和电感组成。LC传感器体积小、成本低、功耗较低,且不需要更换电池,可用于pH监测、温度、湿度、生物电位和应变等参数的检测,有着很重要的使用价值和科研意义。LC passive wireless sensor is essentially an LC resonant loop, which is composed of passive components such as resistors, capacitors and inductors. LC sensor is small in size, low in cost, low in power consumption, and does not require battery replacement. It can be used for pH monitoring, temperature, humidity, biopotential and strain detection, etc., and has very important use value and scientific research significance.

然而传感器往往会由于外界环境变化(温度,电磁干扰等),而出现漂移现象,即传感器响应偏离基准值,这会导致后续的模式识别结果不准确。本发明针对LC湿度传感器因温度变化造成的响应数据不稳定问题,采用神经网络模型对LC湿度传感器的温度漂移进行补偿,对获得更稳定、更可靠的输出结果具有重要意义。However, the sensor often drifts due to changes in the external environment (temperature, electromagnetic interference, etc.), that is, the sensor response deviates from the reference value, which will lead to inaccurate subsequent pattern recognition results. The present invention aims to solve the problem of unstable response data of LC humidity sensor caused by temperature change, and adopts a neural network model to compensate for the temperature drift of LC humidity sensor, which is of great significance for obtaining more stable and reliable output results.

有鉴于此,有必要设计一种基于机器学习的LC湿度传感器漂移补偿方法,以解决上述问题。In view of this, it is necessary to design a LC humidity sensor drift compensation method based on machine learning to solve the above problems.

发明内容Summary of the invention

本发明的目的在于针对LC湿度传感器因温度变化造成的响应数据不稳定问题,提供一种基于机器学习的LC湿度传感器漂移补偿方法。The purpose of the present invention is to provide a LC humidity sensor drift compensation method based on machine learning to address the problem of unstable response data of the LC humidity sensor caused by temperature changes.

为实现以上目的,本发明采用以下技术方案,包括以下步骤:To achieve the above purpose, the present invention adopts the following technical solution, including the following steps:

S1、对LC湿度传感器进行温度测试实验,获取原始数据,所述原始数据包括多组数据,每组所述数据包括湿度环境及温度条件下所述LC湿度传感器的响应数据;S1. Perform a temperature test experiment on the LC humidity sensor to obtain raw data, wherein the raw data includes multiple groups of data, each group of data includes response data of the LC humidity sensor under humidity environment and temperature conditions;

S2、将所述LC湿度传感器的响应数据进行预处理,提取响应带宽BW,实部阻抗最大值Re()Max,及所述最大值Re(/>)Max对应的谐振频率Freq,利用小波分析对所述谐振频率Freq降噪,并将降噪后的数据进行归一化处理之后作为特征建立数据集;S2, pre-processing the response data of the LC humidity sensor, extracting the response bandwidth BW, the real impedance maximum value Re ( )Max, and the maximum value Re(/> ) The resonant frequency Freq corresponding to Max, the resonant frequency Freq is denoised by wavelet analysis, and the denoised data is normalized and used as a feature to establish a data set;

S3、初始化BP神经网络,采用遗传算法优化BP神经网络模型,得到GA-BP模型;S3, initialize the BP neural network, use genetic algorithm to optimize the BP neural network model, and obtain the GA-BP model;

S4、用训练集对所述GA-BP模型进行训练,将训练好的模型用于漂移数据测试集的补偿。S4. The GA-BP model is trained using the training set, and the trained model is used to compensate for the drift data test set.

作为本发明的进一步改进,所述S2中的所述小波分析所采用的基函数为db10,分解层数为3层。As a further improvement of the present invention, the basis function used in the wavelet analysis in S2 is db10, and the number of decomposition levels is 3.

作为本发明的进一步改进,所述S2中的所述小波分析的阈值选取为改进的固定阈值,表达式为:As a further improvement of the present invention, the threshold of the wavelet analysis in S2 is selected as an improved fixed threshold, and the expression is:

,

阈值函数选取为改进的阈值函数,所述阈值函数的公式为:The threshold function is selected as an improved threshold function, and the formula of the threshold function is:

,

其中为信号标准方差,N为数据长度,j为分解层数,/>为估计小波系数,为分解后的小波系数,sgn(*)为符号分段函数。in is the signal standard deviation, N is the data length, j is the number of decomposition layers, /> To estimate the wavelet coefficients, is the decomposed wavelet coefficient, sgn(*) is the symbolic piecewise function.

作为本发明的进一步改进,所述S3中的所述BP神经网络包括1个输入层、2个隐藏层和1个输出层,所述输入层有3个输入,所述2个隐藏层分别有m和n个神经元。As a further improvement of the present invention, the BP neural network in S3 includes 1 input layer, 2 hidden layers and 1 output layer, the input layer has 3 inputs, and the 2 hidden layers have m and n neurons respectively.

作为本发明的进一步改进,所述BP神经网络是通过输入的线性组合,然后通过神经元激活函数进行非线性变换,所述神经元激活函数选用Sigmoid函数,将估计值和实际测量值的均方误差MSE函数作为误差函数,并利用反向传播算法最小化所述误差函数,其中,所述均方误差MSE函数的表达式为:As a further improvement of the present invention, the BP neural network is a linear combination of inputs, and then a nonlinear transformation is performed through a neuron activation function, the neuron activation function uses a Sigmoid function, and the mean square error MSE function of the estimated value and the actual measured value is used as the error function, and the back propagation algorithm is used to minimize the error function, wherein the expression of the mean square error MSE function is:

其中,为实际观测值, />为模型预测输出值,N为测试点数。in, is the actual observed value, /> is the model's predicted output value, and N is the number of test points.

作为本发明的进一步改进,所述S3中的所述遗传算法优化BP神经网络,通过不断地进化种群中的个体,通过选择、交叉和变异探索超参数空间,得到最佳的超参数组合,多次迭代后,所述遗传算法收敛到具有高适应度的个体,所述个体对应于所述超参数组合。As a further improvement of the present invention, the genetic algorithm in S3 optimizes the BP neural network by continuously evolving individuals in the population and exploring the hyperparameter space through selection, crossover and mutation to obtain the best hyperparameter combination. After multiple iterations, the genetic algorithm converges to individuals with high fitness, which correspond to the hyperparameter combination.

作为本发明的进一步改进,所述遗传算法优化所述BP神经网络时,使用决定系数作为模型评价指标,决定系数/>的表达式为:As a further improvement of the present invention, when the genetic algorithm optimizes the BP neural network, the determination coefficient is used As a model evaluation index, the determination coefficient/> The expression is:

其中,为实际观测值,/>为实际观测值的平均值,/>为模型预测输出值,N为测试点数。in, is the actual observed value, /> is the average value of the actual observations, /> is the model's predicted output value, and N is the number of test points.

本发明的有益效果为:The beneficial effects of the present invention are:

本发明通过神经网络模型对LC湿度传感器温度漂移进行补偿,经过对温度漂移数据集的学习,该模型系数达到0.966,表明模型能解释96.6%的不确定性,取得了良好的效果。The present invention uses a neural network model to compensate for the temperature drift of the LC humidity sensor. After learning the temperature drift data set, the model The coefficient reaches 0.966, indicating that the model can explain 96.6% of the uncertainty and achieves good results.

附图说明BRIEF DESCRIPTION OF THE DRAWINGS

图1是本发明实施例使用神经网络模型的LC无源无线传感器温度漂移抑制方法流程图;FIG1 is a flow chart of a method for suppressing temperature drift of an LC passive wireless sensor using a neural network model according to an embodiment of the present invention;

图2是建立LC无源无线湿度传感器温度漂移数据集选取的特征图;FIG2 is a feature diagram selected to establish a temperature drift data set for a LC passive wireless humidity sensor;

图3对温度为15℃,相对湿度从50%RH到95%RH变化时的实部阻抗最大值数据进行降噪效果图;FIG3 is a diagram showing the noise reduction effect of the real impedance maximum value data when the temperature is 15°C and the relative humidity changes from 50%RH to 95%RH;

图4是对温度为35℃,相对湿度从50%RH到95%RH变化时的谐振频率数据进行降噪效果图;FIG4 is a diagram showing the noise reduction effect of the resonant frequency data when the temperature is 35°C and the relative humidity changes from 50%RH to 95%RH;

图5是本发明设计的初始化神经网络图;FIG5 is a diagram of an initialization neural network designed by the present invention;

图6是本发明提出的用训练好的神经网络模型对温度漂移数据拟合结果与真实结果对比分析图。FIG6 is a comparative analysis diagram of the temperature drift data fitting results and the actual results using the trained neural network model proposed by the present invention.

具体实施方式Detailed ways

为了使本发明的目的、技术方案和优点更加清楚,下面结合附图和具体实施例对本发明进行详细描述。In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

在此,还需要说明的是,为了避免因不必要的细节而模糊了本发明,在附图中仅仅示出了与本发明的方案密切相关的结构和/或处理步骤,而省略了与本发明关系不大的其他细节。It should also be noted that, in order to avoid obscuring the present invention due to unnecessary details, only structures and/or processing steps closely related to the scheme of the present invention are shown in the drawings, while other details that are not closely related to the present invention are omitted.

如图1所示,为本发明的一种基于神经网络模型的一种基于机器学习的LC湿度传感器漂移补偿方法步骤示意图,所述方法包括:As shown in FIG1 , it is a schematic diagram of the steps of a method for drift compensation of an LC humidity sensor based on a neural network model and machine learning according to the present invention, the method comprising:

S1、对LC无源无线湿度传感器进行温度测试实验,获取原始数据,原始数据包括多组数据,每组数据包括湿度环境及温度条件下LC湿度传感器响应数据。S1. Conduct a temperature test experiment on the LC passive wireless humidity sensor to obtain raw data. The raw data includes multiple groups of data, each group of data includes the response data of the LC humidity sensor under humidity environment and temperature conditions.

S2、将LC响应数据进行预处理,图2为温度为25℃、相对湿度为60%RH时LC湿度传感器的响应曲线,如图2所示提取响应带宽BW,实部阻抗最大值Re()Max,及其最大值时对应的谐振频率Freq。S2. Preprocess the LC response data. Figure 2 shows the response curve of the LC humidity sensor when the temperature is 25°C and the relative humidity is 60%RH. As shown in Figure 2, the response bandwidth BW and the maximum value of the real impedance Re ( )Max, and the corresponding resonant frequency Freq when it reaches its maximum value.

利用小波分析对提取的数据进行降噪,采用的基函数为db10,分解层数为3层,小波分析阈值的选取为改进的固定阈值,表达式为:Wavelet analysis is used to reduce noise on the extracted data. The basis function used is db10, the number of decomposition layers is 3, and the wavelet analysis threshold is selected as an improved fixed threshold. The expression is:

,

阈值函数选取为改进的阈值函数,所述阈值函数的公式为:The threshold function is selected as an improved threshold function, and the formula of the threshold function is:

,

其中为信号标准方差,N为数据长度,j为分解层数,/>为估计小波系数,为分解后的小波系数,sgn(*)为符号分段函数。in is the signal standard deviation, N is the data length, j is the number of decomposition layers, /> To estimate the wavelet coefficients, is the decomposed wavelet coefficient, sgn(*) is the signed piecewise function.

图3是对温度为15℃,相对湿度从50%RH到95%RH变化时的实部阻抗最大值数据进行降噪,图4是对温度为35℃,相对湿度从50%RH到95%RH变化时的谐振频率数据进行降噪,将降噪后的数据进行归一化处理之后作为特征建立数据集。Figure 3 shows the denoising of the real impedance maximum value data when the temperature is 15°C and the relative humidity changes from 50%RH to 95%RH. Figure 4 shows the denoising of the resonant frequency data when the temperature is 35°C and the relative humidity changes from 50%RH to 95%RH. The denoised data are normalized and used as features to establish a data set.

S3、初始化BP神经网络,如图5,包括1个输入层,2个隐藏层和1个输出层,其输入层有3个输入,2个隐藏层分别有m和n个神经元。神经网络是通过输入的线性组合,然后通过激活函数进行非线性变换来构造的,神经元激活函数选用Sigmoid函数,将估计值和实际测量值的均方误差(MSE)函数作为误差函数,并利用反向传播算法最小化误差函数,其中均方误差的表达式为:S3. Initialize the BP neural network, as shown in Figure 5, including 1 input layer, 2 hidden layers and 1 output layer. The input layer has 3 inputs, and the 2 hidden layers have m and n neurons respectively. The neural network is constructed by linear combination of inputs and then nonlinear transformation through activation function. The neuron activation function uses Sigmoid function, and the mean square error ( MSE ) function of the estimated value and the actual measured value is used as the error function. The back propagation algorithm is used to minimize the error function, where the expression of the mean square error is:

其中,为实际观测值, />为模型预测输出值,N为测试点数。in, is the actual observed value, /> is the model's predicted output value, and N is the number of test points.

采用遗传算法优化BP神经网络模型,得到GA-BP模型对网络的超参数进行优化,寻找最优学习率和隐藏层的神经元个数的组合。The genetic algorithm is used to optimize the BP neural network model, and the GA-BP model is obtained to optimize the hyperparameters of the network and find the optimal combination of learning rate and the number of neurons in the hidden layer.

使用决定系数()作为模型评价指标,/>系数反映因变量的全部变异能通过回归关系被自变量解释的比例,/>系数越接近1,回归拟合的效果越好,其表达式为:Using the coefficient of determination ( ) as the model evaluation index, /> The coefficient reflects the proportion of the total variation of the dependent variable that can be explained by the independent variable through the regression relationship. The closer the coefficient is to 1, the better the regression fitting effect is. Its expression is:

其中,为实际观测值,/>为实际观测值的平均值,/>为模型预测输出值,N为测试点数。in, is the actual observed value, /> is the average value of the actual observations, /> is the model's predicted output value, and N is the number of test points.

该模型使用的优化器为Adam,迭代次数(epoch)为10000次,所使用的超参数包括第一层隐藏层的神经元个数为95,第二层隐藏层神经元的个数为81,学习率为0.00095。The optimizer used in this model is Adam, the number of iterations (epochs) is 10,000, the hyperparameters used include 95 neurons in the first hidden layer, 81 neurons in the second hidden layer, and a learning rate of 0.00095.

S4、用训练集对所述GA-BP模型进行训练,将训练好的模型用于漂移数据测试集的补偿。图6为将训练好的神经网络模型对温度漂移数据拟合结果与真实结果对比分析图,系数达到0.966。S4, train the GA-BP model with the training set, and use the trained model to compensate the drift data test set. Figure 6 is a comparative analysis of the fitting results of the trained neural network model to the temperature drift data and the actual results, The coefficient reaches 0.966.

综上所述,本发明通过神经网络模型对LC湿度传感器温度漂移进行补偿,经过对温度漂移数据集的学习,该模型系数达到0.966,表明模型能解释96.6%的不确定性,取得了良好的效果。In summary, the present invention compensates the temperature drift of the LC humidity sensor through a neural network model. After learning the temperature drift data set, the model The coefficient reaches 0.966, indicating that the model can explain 96.6% of the uncertainty and achieves good results.

以上实施例仅用以说明本发明的技术方案而非限制,尽管参照较佳实施例对本发明进行了详细说明,本领域的普通技术人员应当理解,可以对本发明的技术方案进行修改或者等同替换,而不脱离本发明技术方案的精神和范围。The above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims (7)

1.一种基于机器学习的LC湿度传感器漂移补偿方法,其特征在于,包括:1. A method for compensating drift of an LC humidity sensor based on machine learning, comprising: S1、对LC湿度传感器进行温度测试实验,获取原始数据,所述原始数据包括多组数据,每组所述数据包括湿度及温度条件下所述LC湿度传感器的响应数据;S1. Perform a temperature test experiment on the LC humidity sensor to obtain raw data, wherein the raw data includes multiple groups of data, each group of data includes response data of the LC humidity sensor under humidity and temperature conditions; S2、将所述LC湿度传感器的响应数据进行预处理,提取响应带宽BW,实部阻抗最大值Re()Max,及所述最大值Re(/>)Max对应的谐振频率Freq,利用小波分析对所述谐振频率Freq降噪,并将降噪后的数据进行归一化处理作为特征建立数据集;S2, pre-processing the response data of the LC humidity sensor, extracting the response bandwidth BW, the maximum value of the real impedance Re ( )Max, and the maximum value Re(/> ) The resonant frequency Freq corresponding to Max, the resonant frequency Freq is denoised by wavelet analysis, and the denoised data is normalized as a feature to establish a data set; S3、初始化BP神经网络,采用遗传算法优化BP神经网络模型,得到GA-BP模型对网络的超参数进行优化,寻找最优学习率和隐藏层的神经元个数的组合;S3, initialize the BP neural network, use genetic algorithm to optimize the BP neural network model, obtain the GA-BP model to optimize the network's hyperparameters, and find the combination of the optimal learning rate and the number of neurons in the hidden layer; S4、所述数据集包括训练集和测试集,用所述训练集对所述GA-BP模型进行训练,将训练好的模型用于漂移数据的所述测试集的补偿;S4, the data set includes a training set and a test set, the training set is used to train the GA-BP model, and the trained model is used to compensate for the drift data of the test set; 所述S3中的所述遗传算法优化BP神经网络,通过不断地进化种群中的个体,通过选择、交叉和变异探索超参数空间,得到最佳的超参数组合,多次迭代后,所述遗传算法收敛到具有高适应度的个体,所述个体对应于所述超参数组合。The genetic algorithm in S3 optimizes the BP neural network by continuously evolving individuals in the population and exploring the hyperparameter space through selection, crossover and mutation to obtain the best hyperparameter combination. After multiple iterations, the genetic algorithm converges to individuals with high fitness, which correspond to the hyperparameter combination. 2.根据权利要求1所述的一种基于机器学习的LC湿度传感器漂移补偿方法,其特征在于,所述S2中的所述小波分析所采用的基函数为db10,信号的分解层数为3层。2. The LC humidity sensor drift compensation method based on machine learning according to claim 1 is characterized in that the basis function used in the wavelet analysis in S2 is db10, and the number of signal decomposition layers is 3. 3.根据权利要求2所述的一种基于机器学习的LC湿度传感器漂移补偿方法,其特征在于,所述S2中的所述小波分析的阈值选取为改进的固定阈值,表达式为:3. The LC humidity sensor drift compensation method based on machine learning according to claim 2 is characterized in that the threshold of the wavelet analysis in S2 is selected as an improved fixed threshold, and the expression is: , 阈值函数选取为改进的阈值函数,所述阈值函数的公式为:The threshold function is selected as an improved threshold function, and the formula of the threshold function is: , 其中为信号标准方差,N为数据长度,j为分解层数,/>为估计小波系数,/>为分解后的小波系数,sgn(*)为符号分段函数。in is the signal standard deviation, N is the data length, j is the number of decomposition layers, /> To estimate the wavelet coefficients, /> is the decomposed wavelet coefficient, sgn(*) is the signed piecewise function. 4.根据权利要求1所述的一种基于机器学习的LC湿度传感器漂移补偿方法,其特征在于,所述S3中的所述BP神经网络包括1个输入层、2个隐藏层和1个输出层,所述输入层有3个输入,所述2个隐藏层分别有m和n个神经元。4. The LC humidity sensor drift compensation method based on machine learning according to claim 1, characterized in that the BP neural network in S3 includes 1 input layer, 2 hidden layers and 1 output layer, the input layer has 3 inputs, and the 2 hidden layers have m and n neurons respectively. 5.根据权利要求4所述的一种基于机器学习的LC湿度传感器漂移补偿方法,其特征在于,所述BP神经网络是通过输入的线性组合,然后通过神经元激活函数进行非线性变换,所述神经元激活函数选用Sigmoid函数,将估计值和实际测量值的均方误差MSE函数作为误差函数,并利用反向传播算法最小化所述误差函数,其中,所述均方误差MSE函数的表达式为:5. A method for drift compensation of LC humidity sensor based on machine learning according to claim 4, characterized in that the BP neural network is a linear combination of inputs, and then a nonlinear transformation is performed through a neuron activation function, the neuron activation function selects a Sigmoid function, and the mean square error MSE function of the estimated value and the actual measured value is used as the error function, and the back propagation algorithm is used to minimize the error function, wherein the expression of the mean square error MSE function is: 其中,/>为实际观测值, />为模型预测输出值,n为测试点数。 Among them,/> is the actual observed value, /> is the model's predicted output value, and n is the number of test points. 6.根据权利要求1所述的一种基于机器学习的LC湿度传感器漂移补偿方法,其特征在于,所述遗传算法优化所述BP神经网络时,使用决定系数作为模型评价指标,决定系数的表达式为:6. The LC humidity sensor drift compensation method based on machine learning according to claim 1, characterized in that when the genetic algorithm optimizes the BP neural network, the determination coefficient is used As a model evaluation indicator, the coefficient of determination The expression is: 其中,/>为实际观测值,/>为实际观测值的平均值,为模型预测输出值,n为测试点数。 Among them,/> is the actual observed value, /> is the average of the actual observed values, is the model's predicted output value, and n is the number of test points. 7.根据权利要求1所述的一种基于机器学习的LC湿度传感器漂移补偿方法,其特征在于,所述BP神经网络模型使用Adam优化器优化模型参数,迭代次数为10000次。7. The LC humidity sensor drift compensation method based on machine learning according to claim 1 is characterized in that the BP neural network model uses an Adam optimizer to optimize model parameters, and the number of iterations is 10,000 times.
CN202410153167.5A 2024-02-04 2024-02-04 A drift compensation method for LC humidity sensor based on machine learning Active CN117686555B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410153167.5A CN117686555B (en) 2024-02-04 2024-02-04 A drift compensation method for LC humidity sensor based on machine learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410153167.5A CN117686555B (en) 2024-02-04 2024-02-04 A drift compensation method for LC humidity sensor based on machine learning

Publications (2)

Publication Number Publication Date
CN117686555A CN117686555A (en) 2024-03-12
CN117686555B true CN117686555B (en) 2024-05-14

Family

ID=90126838

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410153167.5A Active CN117686555B (en) 2024-02-04 2024-02-04 A drift compensation method for LC humidity sensor based on machine learning

Country Status (1)

Country Link
CN (1) CN117686555B (en)

Citations (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5554273A (en) * 1995-07-26 1996-09-10 Praxair Technology, Inc. Neural network compensation for sensors
CN102914623A (en) * 2012-10-19 2013-02-06 南京信息工程大学 Fusing method of temperature compensation of humidity sensor
CN103593538A (en) * 2013-11-28 2014-02-19 东南大学 Fiber optic gyroscope temperature drift modeling method by optimizing dynamic recurrent neural network through genetic algorithm
WO2015143127A1 (en) * 2014-03-19 2015-09-24 Actiwave Ab Non-linear control of loudspeakers
CN105223241A (en) * 2015-09-18 2016-01-06 南京信息工程大学 A kind of compensation method of humidity sensor
CN106991491A (en) * 2017-02-21 2017-07-28 南京邮电大学 A kind of environment pollution monitoring system of the RBF neural based on genetic algorithm optimization
CN109297533A (en) * 2018-09-20 2019-02-01 北京恩福健康科技有限公司 A kind of method of precise measurement skin surface temperature and humidity
CN109666717A (en) * 2018-12-06 2019-04-23 浙江大学 The prediction technique of degree is infected by aspergillus fungi based on the rice of electronic nose
CN111163011A (en) * 2020-01-19 2020-05-15 烟台持久钟表有限公司 Data processing method of wireless router based on PTP (precision time protocol)
CN112846938A (en) * 2021-01-05 2021-05-28 北京信息科技大学 Main shaft rotation precision degradation traceability system under cutting working condition
CN115049019A (en) * 2022-07-25 2022-09-13 湖南工商大学 Method and device for evaluating arsenic adsorption performance of metal organic framework and related equipment
CN115307689A (en) * 2022-10-09 2022-11-08 中国石油大学(华东) Wireless passive temperature and humidity monitoring device and method
KR20230075862A (en) * 2021-11-23 2023-05-31 유엔젤주식회사 Automatic machine learning-based complex environment analysis artificial intelligence model automatic distribution management system and its operation method
CN116307034A (en) * 2022-12-02 2023-06-23 中国石油大学(华东) Oil production or CO production based on discrete wavelet transformation and neural network 2 Buried quantity prediction method
WO2023168371A2 (en) * 2022-03-02 2023-09-07 Fdh Infrastructure Services, Llc Machine learning to predict bolt tension
CN116720620A (en) * 2023-06-10 2023-09-08 河南工业大学 A method for predicting ventilation temperature of stored grains based on IPSO algorithm optimized CNN-BiGRU-Attention network model
WO2023231204A1 (en) * 2022-06-02 2023-12-07 苏州大学 Sensor physical quantity regression method based on ics-bp neural network
CN117245872A (en) * 2023-10-08 2023-12-19 东莞市坤裕精密模具有限公司 State compensation model control method and system for batch injection molding process
WO2024001719A1 (en) * 2022-06-30 2024-01-04 西南交通大学 Parameter optimization method and system for vibration compaction of high-speed rail filler

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070035114A1 (en) * 1992-05-05 2007-02-15 Automotive Technologies International, Inc. Device and Method for Deploying a Vehicular Occupant Protection System
EP2520014A1 (en) * 2009-12-28 2012-11-07 Flyback Energy, Inc. Controllable universal power supply with reactive power management
US9696272B2 (en) * 2015-08-07 2017-07-04 Silicon Laboratories Inc. Systems and methods for humidity measurement using dielectric material-based relative humidity sensors
US11533527B2 (en) * 2018-05-09 2022-12-20 Pluto Inc. Methods and systems for generating and providing program guides and content
KR102364019B1 (en) * 2020-08-12 2022-02-18 울산과학기술원 Sensor drift compensation method and device

Patent Citations (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5554273A (en) * 1995-07-26 1996-09-10 Praxair Technology, Inc. Neural network compensation for sensors
CN102914623A (en) * 2012-10-19 2013-02-06 南京信息工程大学 Fusing method of temperature compensation of humidity sensor
CN103593538A (en) * 2013-11-28 2014-02-19 东南大学 Fiber optic gyroscope temperature drift modeling method by optimizing dynamic recurrent neural network through genetic algorithm
WO2015143127A1 (en) * 2014-03-19 2015-09-24 Actiwave Ab Non-linear control of loudspeakers
CN105223241A (en) * 2015-09-18 2016-01-06 南京信息工程大学 A kind of compensation method of humidity sensor
CN106991491A (en) * 2017-02-21 2017-07-28 南京邮电大学 A kind of environment pollution monitoring system of the RBF neural based on genetic algorithm optimization
CN109297533A (en) * 2018-09-20 2019-02-01 北京恩福健康科技有限公司 A kind of method of precise measurement skin surface temperature and humidity
CN109666717A (en) * 2018-12-06 2019-04-23 浙江大学 The prediction technique of degree is infected by aspergillus fungi based on the rice of electronic nose
CN111163011A (en) * 2020-01-19 2020-05-15 烟台持久钟表有限公司 Data processing method of wireless router based on PTP (precision time protocol)
CN112846938A (en) * 2021-01-05 2021-05-28 北京信息科技大学 Main shaft rotation precision degradation traceability system under cutting working condition
KR20230075862A (en) * 2021-11-23 2023-05-31 유엔젤주식회사 Automatic machine learning-based complex environment analysis artificial intelligence model automatic distribution management system and its operation method
WO2023168371A2 (en) * 2022-03-02 2023-09-07 Fdh Infrastructure Services, Llc Machine learning to predict bolt tension
WO2023231204A1 (en) * 2022-06-02 2023-12-07 苏州大学 Sensor physical quantity regression method based on ics-bp neural network
WO2024001719A1 (en) * 2022-06-30 2024-01-04 西南交通大学 Parameter optimization method and system for vibration compaction of high-speed rail filler
CN115049019A (en) * 2022-07-25 2022-09-13 湖南工商大学 Method and device for evaluating arsenic adsorption performance of metal organic framework and related equipment
CN115307689A (en) * 2022-10-09 2022-11-08 中国石油大学(华东) Wireless passive temperature and humidity monitoring device and method
CN116307034A (en) * 2022-12-02 2023-06-23 中国石油大学(华东) Oil production or CO production based on discrete wavelet transformation and neural network 2 Buried quantity prediction method
CN116720620A (en) * 2023-06-10 2023-09-08 河南工业大学 A method for predicting ventilation temperature of stored grains based on IPSO algorithm optimized CNN-BiGRU-Attention network model
CN117245872A (en) * 2023-10-08 2023-12-19 东莞市坤裕精密模具有限公司 State compensation model control method and system for batch injection molding process

Non-Patent Citations (7)

* Cited by examiner, † Cited by third party
Title
A GA-BP Neural Network Regression Model for Predicting Soil Moisture in Slope Ecological Protection;Liu, DW等;《SUSTAINABILITY》;20220211;第14卷(第3期);第362-374页 *
Modeling and Analysis of Adaptive Temperature Compensation for Humidity Sensors;Xu, W等;《ELECTRONICS》;20190412;第8卷(第4期);第1133-1139页 *
基于改进GA-BP神经网络的湿度传感器的温度补偿;彭基伟等;《仪器仪表学报》;20130115;第34卷(第1期);第153-160页 *
改进AFSA-BP神经网络的湿度传感器温度补偿研究;何怡刚等;《电子测量与仪器学报》;20180715;第32卷(第7期);第95-100页 *
无源无线温湿度传感器研究;任青;《中国博士学位论文全文数据库信息科技辑》;20170915(第9期);第I140-37页 *
电子鼻传感器的在线漂移补偿研究;和峥;《中国优秀硕士学位论文全文数据库信息科技辑》;20190415(第4期);第I140-147页 *
集成LC无源无线传感的微流控液滴表征系统;王蜜;《中国优秀硕士学位论文全文数据库信息科技辑》;20220615(第6期);第I135-190 *

Also Published As

Publication number Publication date
CN117686555A (en) 2024-03-12

Similar Documents

Publication Publication Date Title
CN110472817B (en) XGboost integrated credit evaluation system and method combined with deep neural network
CN110232341B (en) Semi-supervised learning image identification method based on convolution-stacking noise reduction coding network
CN110213244A (en) A kind of network inbreak detection method based on space-time characteristic fusion
Zhang et al. A novel semi-supervised learning approach in artificial olfaction for E-nose application
CN113159389A (en) Financial time sequence prediction method based on deep forest generation countermeasure network
CN111177224B (en) An Unsupervised Anomaly Detection Method for Time Series Based on Conditional Normalized Flow Model
CN108009571A (en) A kind of semi-supervised data classification method of new direct-push and system
CN112766355A (en) Electroencephalogram signal emotion recognition method under label noise
CN111598167B (en) Small sample image identification method and system based on graph learning
CN116897346A (en) Structured graph convolution network with random mask for network embedding
CN114330572A (en) Anomaly detection method, system and computer storage medium based on contrastive learning
CN114154557A (en) Cancer tissue classification method, device, electronic device and storage medium
CN114027786A (en) Sleep-disordered breathing detection method and system based on self-supervised memory network
CN117034030A (en) Electroencephalo-gram data alignment algorithm based on positive and negative two-way information fusion
CN111898799B (en) A Power Load Forecasting Method Based on BFA-Elman
CN117371573A (en) Time sequence prediction method, device and medium based on TrAdaBoost-LSTM
CN117686555B (en) A drift compensation method for LC humidity sensor based on machine learning
CN117606801A (en) A cross-domain bearing fault diagnosis method based on multi-representation adaptive network
Shen et al. Dual separated attention-based graph neural network
Ukil et al. Sig-R 2 ResNet: residual network with signal processing-refined residual mapping, auto-tuned L 1-regularization with modified Adam optimizer for time series classification
Yang et al. A structure optimization algorithm of neural networks for large-scale data sets
CN118747302A (en) A method and system for evaluating the state of electric energy meters based on coefficient of variation method and K-means
CN116627035A (en) Internet of things intelligent monitoring method and system for livestock and poultry housing cultivation environment with big data
Liu et al. GCD: Graph contrastive denoising module for GNNs in EEG classification
Zhang et al. Intrinsic plasticity based inference acceleration for spiking multi-layer perceptron

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant