CN111474297B - Online drift compensation method for sensor in bionic olfaction system - Google Patents

Online drift compensation method for sensor in bionic olfaction system Download PDF

Info

Publication number
CN111474297B
CN111474297B CN202010158546.5A CN202010158546A CN111474297B CN 111474297 B CN111474297 B CN 111474297B CN 202010158546 A CN202010158546 A CN 202010158546A CN 111474297 B CN111474297 B CN 111474297B
Authority
CN
China
Prior art keywords
samples
batch
sample
sensor
source domain
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
CN202010158546.5A
Other languages
Chinese (zh)
Other versions
CN111474297A (en
Inventor
陶洋
杨皓诚
梁志芳
黎春燕
孔宇航
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.)
Chongqing University of Post and Telecommunications
Original Assignee
Chongqing University of Post 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 Chongqing University of Post and Telecommunications filed Critical Chongqing University of Post and Telecommunications
Priority to CN202010158546.5A priority Critical patent/CN111474297B/en
Publication of CN111474297A publication Critical patent/CN111474297A/en
Application granted granted Critical
Publication of CN111474297B publication Critical patent/CN111474297B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
    • G01N33/0001—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00 by organoleptic means
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
    • G01N33/0004—Gaseous mixtures, e.g. polluted air
    • G01N33/0006—Calibrating gas analysers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
    • G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification

Abstract

The invention relates to an online drift compensation method for a sensor in a bionic olfactory system, and belongs to the technical field of sensors. The method comprises the following steps: step 1) performing source domain reconstruction according to an input sample batch number; and 2) constructing a classification model by using the reconstructed source domain and target domain samples and storing a prediction result. The method comprises the steps of utilizing output response samples of sensors in a bionic olfactory system in two successive batches, namely performing source domain reconstruction on samples of a previous batch which are predicted through a classification model and samples of an initial batch which are artificially marked, and then building a classification model through condition distribution self-adaption and manifold regularization to realize online drift compensation of the sensors in the bionic olfactory system. The gas identification model can be continuously updated along with the drift of the sensor, so that the method is more suitable for actual production and use scenes of the bionic olfactory system in a real scene, and the service life of equipment can be prolonged.

Description

Online drift compensation method for sensor in bionic olfaction system
Technical Field
The invention belongs to the technical field of sensors, and relates to an online drift compensation method for a sensor in a bionic olfaction system.
Background
The bionic olfactory system consists of a gas sensor array, a signal preprocessing unit and a mode recognition algorithm and can be used for gas recognition. After the gas is introduced into the system, the sensor array generates corresponding electric signal response according to the gas characteristics, and the preprocessed signal is converted into a gas recognition result through a pattern recognition algorithm.
The sensor drifts due to its aging or gas poisoning. The drift problem has long been associated with the use of biomimetic olfactory systems and cannot be avoided. Drift can change the output response of the sensor, and further the classification model established initially can not accurately predict the samples collected in the later period.
In recent years, many algorithms for sensor drift compensation are proposed, and mainly include three types, namely signal preprocessing, component correction and machine learning, and although the algorithms can realize the drift compensation of the sensor to a certain extent, most of the algorithms belong to off-line methods, and require periodic recovery of equipment to complete manual correction, so that the algorithms are not suitable for practical application scenes. In addition, the characteristic distribution of the output response samples before and after the sensor drift is different, the related work of the conventional drift compensation focuses on reducing the edge distribution difference, and the influence caused by the condition distribution difference is not considered.
Therefore, how to reduce the condition distribution difference caused by the drift of the sensor and realize the effective online update of the classification model has great influence on the correctness of the judgment result of the bionic olfactory system gas. The online drift compensation method for the sensor in the bionic olfactory system disclosed by the patent can construct a classification model through condition distribution self-adaption and manifold regularization, meanwhile, online update of the model is realized by utilizing source domain reconstruction, compensation of samples collected by the sensor which has drifted in the bionic olfactory system is completed, and the online drift compensation method is more reasonable in a practical use scene.
Disclosure of Invention
In view of the above, the present invention provides an online drift compensation method for a sensor in a bionic olfactory system, which includes performing source domain reconstruction on output response samples of the sensor in the bionic olfactory system in two consecutive batches, that is, samples of a previous batch that have been predicted by a classification model and samples of an initial batch that have been artificially labeled, and then building a classification model through condition distribution adaptation and manifold regularization, so as to implement online drift compensation for the sensor in the bionic olfactory system.
In order to achieve the purpose, the invention provides the following technical scheme:
an on-line drift compensation method for a sensor in a bionic olfactory system comprises the following steps:
step 1) performing source domain reconstruction according to an input sample batch number;
and 2) constructing a classification model by using the reconstructed source domain and target domain samples and storing a prediction result.
Optionally, the step 1) includes the following steps:
step 11) inputting the batch number a of the sample;
step 12) carrying out source domain reconstruction according to the batch number a, and when a is 1, carrying out source domain DsSelecting an initial lot of labeled sample sets collected when the sensor is not drifting
Figure RE-GDA0002487699360000021
n1Is the number of initial batch samples, otherwise, DsSample set with class prediction completed by previous batch of current target domain
Figure RE-GDA0002487699360000022
And D1Common construction, i.e. Ds＝D1∪DaWherein n isaFor the number of samples in batch a, when a is 1, there is no classification prediction result of the previous batch due to the first model construction, so Ds＝D1The number of reconstructed source domain samples is:
Figure RE-GDA0002487699360000023
optionally, the step 2) includes the following steps:
step 21) inputting unlabeled a +1 batch target domain samples
Figure RE-GDA0002487699360000024
ntIs the number of samples in the target domain;
step 22) Using principal component analysisXsAnd XtDimension reduction to p dimension to generate subspace SsAnd St，Gd×pFor the set of all p-dimensional subspaces, each subspace is considered as the Grassman manifold space Gd×pAt one point above, let Φ (t) be Gd×pA geodesic line of (1), wherein t ∈ [0,1 ]]，Φ(0)＝SsAnd Φ (1) ═ StAs ends of a geodesic line, ziAnd zjIs xiAnd xjThe inner product of the projected feature vector in the infinite dimensional space is expressed as:
Figure RE-GDA0002487699360000025
in the above formula xi,xj∈Ds∪DtAnd G denotes a geodesic core:
Figure RE-GDA0002487699360000026
in the above formula RsFrom XsThe d-p dimensional feature composition, U, remaining after the P dimensional feature extraction by PCA1And U2Are orthogonal matrices and are solved by singular value decomposition1,Λ2And Λ3For a diagonal matrix, the element values in the matrix are respectively:
Figure RE-GDA0002487699360000027
mixing XsAnd XtSample feature space usage after projection
Figure RE-GDA0002487699360000028
And
Figure RE-GDA0002487699360000029
is shown in which
Figure RE-GDA00024876993600000210
Step 23) Using ZsTraining the classifier by k nearest neighbor algorithm with k equal to 1, and then converting Z into ZtBrought into the classifier to obtain pseudo labels
Figure RE-GDA00024876993600000211
Step 24) Using ZsAnd ZtSelecting a Gaussian kernel to construct a kernel function
Figure RE-GDA0002487699360000031
Step 25) determining Z by k-nearest neighbor algorithmsAnd ZtThe neighbor relation of each point in the similarity matrix W is obtained:
Figure RE-GDA0002487699360000032
Figure RE-GDA0002487699360000033
in the above formula r (z)i,zj) 1 represents ziAnd zjAre in a mutual neighbor relationship; after obtaining W, calculating a Laplace matrix L:
Figure RE-GDA0002487699360000034
in the above formula, D is a diagonal matrix consisting of
Figure RE-GDA0002487699360000035
And calculating to obtain, wherein the manifold regular constraint term is expressed as:
Figure RE-GDA0002487699360000036
step 26) for ZsAnd ZtPerforming condition distribution self-adaptation, wherein the condition distribution difference passes through the maximum average value differenceThe measurement is performed on a regenerative Hilbert space, and the statistical estimation is approximated using an empirical estimation:
Figure RE-GDA0002487699360000037
where C represents the total number of classes of labels contained within the sample,
Figure RE-GDA0002487699360000038
consists of two diagonal matrices, wherein:
Figure RE-GDA0002487699360000039
Figure RE-GDA00024876993600000310
Figure RE-GDA00024876993600000311
represents ZsAnd ZtN with the middle label being llSet of classification results for individual samples, f denotes Gd×pThe following classification prediction function:
Figure RE-GDA00024876993600000312
in the above formula, alpha ═ alpha1,α2,...,αN]TFor coefficient vectors, the final calculation formula of the conditional distribution adaptive term is:
MMD2(HK,Qs,Qt)＝tr(fTMf)
in the above formula, each element in M is directly calculated by the following formula:
Figure RE-GDA0002487699360000041
according to the structural risk minimization principle, by combining a manifold regularization term and a condition distribution self-adaptive term, the final optimization target of the classifier f is as follows:
Figure RE-GDA0002487699360000042
in the above equation, U is an indication matrix of the domain where the sample is located, that is:
Figure RE-GDA0002487699360000043
let the partial derivative of α in the optimization equation be 0 to solve for α:
α＝(λ1I+(λ2M+λ3L+U)K)-1UYT
using f to complete a pair
Figure RE-GDA0002487699360000044
Repeating step 26) e times to iteratively update M and α;
step 27) obtaining the prediction label of the sample of the batch
Figure RE-GDA0002487699360000045
And storing the samples to wait for the input of the next batch of samples.
The invention has the beneficial effects that: the method can continuously update the gas identification model along with the drift of the sensor, more accords with the actual production and use scenes of the bionic olfactory system in the actual scene, and can prolong the service life of the equipment.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims thereof.
Drawings
For the purposes of promoting a better understanding of the objects, aspects and advantages of the invention, reference will now be made to the following detailed description taken in conjunction with the accompanying drawings in which:
FIG. 1 is a flow chart of the present invention.
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present invention in a schematic way, and the features in the following embodiments and examples may be combined with each other without conflict.
Wherein the showings are for the purpose of illustrating the invention only and not for the purpose of limiting the same, and in which there is shown by way of illustration only and not in the drawings in which there is no intention to limit the invention thereto; for a better explanation of the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of an actual product; it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
The same or similar reference numerals in the drawings of the embodiments of the present invention correspond to the same or similar components; in the description of the present invention, it should be understood that if there is an orientation or positional relationship indicated by terms such as "upper", "lower", "left", "right", "front", "rear", etc., based on the orientation or positional relationship shown in the drawings, it is only for convenience of description and simplification of description, but it is not an indication or suggestion that the referred device or element must have a specific orientation, be constructed in a specific orientation, and be operated, and therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes, and are not to be construed as limiting the present invention, and the specific meaning of the terms may be understood by those skilled in the art according to specific situations.
The invention provides an online drift compensation method of a sensor in a bionic olfaction system, which comprises the following steps as shown in figure 1:
step 1) performing source domain reconstruction according to an input sample batch number;
further, the step 1) comprises the following steps:
step 11) inputting the batch number a of the sample;
step 12) carrying out source domain reconstruction according to the batch number a, and when a is 1, DsSelecting an initial lot of labeled sample sets collected when the sensor is not drifting
Figure RE-GDA0002487699360000051
n1Is the number of initial batch samples, otherwise, DsSample set with class prediction completed by previous batch of current target domain
Figure RE-GDA0002487699360000052
And D1Common construction, i.e. Ds＝D1∪DaWherein n isaFor the number of samples in batch a, when a is 1, there is no classification prediction result of the previous batch due to the first model construction, so Ds＝D1The number of reconstructed source domain samples is:
Figure RE-GDA0002487699360000053
step 2) constructing a classification model by using the reconstructed source domain and target domain samples and storing a prediction result;
further, the step 2) comprises the following steps:
step 21) inputting unlabeled a +1 batch target domain samples
Figure RE-GDA00024876993600000612
ntIs the number of samples in the target domain;
step 22) analysis of X using principal ComponentssAnd XtDimension reduction to p dimension to generate subspace SsAnd St，Gd×pFor the set of all p-dimensional subspaces, each subspace can be regarded as a Grassmann manifold space Gd×pAt one point above, let Φ (t) be Gd×pA geodesic line of (1), wherein t ∈ [0,1 ]]，Φ(0)＝SsAnd Φ (1) ═ StAs ends of a geodesic line, ziAnd zjIs xiAnd xjThe inner product of the projected feature vector in the infinite dimensional space can be expressed as:
Figure RE-GDA0002487699360000061
in the above formula xi,xj∈Ds∪DtAnd G denotes a geodesic core:
Figure RE-GDA0002487699360000062
in the above formula RsFrom XsThe remaining d-p dimensional features after PCA extraction of the p dimensional features are composed of U1And U2Are orthogonal matrices and can be obtained by singular value decomposition1,Λ2And Λ3The diagonal matrix is formed by the following element values:
Figure RE-GDA0002487699360000063
mixing XsAnd XtSample feature space usage after projection
Figure RE-GDA0002487699360000064
And
Figure RE-GDA0002487699360000065
is shown in which
Figure RE-GDA0002487699360000066
Step 23) Using ZsTraining by k-nearest neighbor algorithm with k-1Classifier, then Z istBrought into the classifier to obtain pseudo labels
Figure RE-GDA0002487699360000067
Step 24) Using ZsAnd ZtSelecting Gaussian kernels to construct kernel functions
Figure RE-GDA0002487699360000068
Step 25) determining Z by k-nearest neighbor algorithmsAnd ZtThe neighbor relation of each point in the similarity matrix W is obtained:
Figure RE-GDA0002487699360000069
Figure RE-GDA00024876993600000610
in the above formula r (z)i,zj) 1 represents ziAnd zjAre in a mutual neighbor relationship. Obtaining W, then calculating a Laplace matrix L:
Figure RE-GDA00024876993600000611
in the above formula, D is a diagonal matrix consisting of
Figure RE-GDA0002487699360000071
And calculating to obtain, wherein the manifold regular constraint term can be expressed as:
Figure RE-GDA0002487699360000072
step 26) for ZsAnd ZtPerforming conditional distribution adaptation, measuring the conditional distribution difference on the regenerated Hilbert space through the maximum mean difference, and approximating statistics by using an empirical estimation formulaEstimating:
Figure RE-GDA0002487699360000073
where C represents the total number of classes of labels contained within the sample,
Figure RE-GDA0002487699360000074
consists of two diagonal matrices, wherein:
Figure RE-GDA0002487699360000075
Figure RE-GDA0002487699360000076
Figure RE-GDA0002487699360000077
represents ZsAnd ZtN with the middle label being llSet of classification results for individual samples, f denotes Gd×pThe following classification prediction functions:
Figure RE-GDA0002487699360000078
in the above formula, alpha ═ alpha1,α2,...,αN]TFor coefficient vectors, the final calculation formula of the conditional distribution adaptive term is:
MMD2(HK,Qs,Qt)＝tr(fTMf)
in the above formula, each element in M can be directly calculated by the following formula:
Figure RE-GDA0002487699360000079
according to the structural risk minimization principle, by combining a manifold regularization term and a condition distribution self-adaptive term, the final optimization target of the classifier f is as follows:
Figure RE-GDA0002487699360000081
in the above equation, U is an indication matrix of the domain where the sample is located, that is:
Figure RE-GDA0002487699360000082
let the partial derivative of α in the optimization equation be 0 to solve for α:
α＝(λ1I+(λ2M+λ3L+U)K)-1UYT
using f to complete pairs
Figure RE-GDA0002487699360000083
Repeating step 26) e times to iteratively update M and α;
step 27) obtaining the prediction label of the sample of the batch
Figure RE-GDA0002487699360000084
And storing the samples to wait for the input of the next batch of samples.
Finally, the above embodiments are only intended to illustrate the technical solutions of the present invention and not to limit the present invention, and although the present invention has been described in detail with reference to the preferred embodiments, it will be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions, and all of them should be covered by the claims of the present invention.

Claims (1)

1. An on-line drift compensation method of a sensor in a bionic olfactory system is characterized in that: the method comprises the following steps:
step 1) performing source domain reconstruction according to an input sample batch number;
step 2) constructing a classification model by using the reconstructed source domain and target domain samples and storing a prediction result;
the step 1) comprises the following steps:
step 11) inputting the batch number a of the sample;
step 12) carrying out source domain reconstruction according to the batch number a, and when a is 1, DsSelecting an initial lot of labeled sample sets collected when the sensor is not drifting
Figure FDA0003568088060000011
n1Is the number of initial batch samples, otherwise, DsSample set with class prediction completed by previous batch of current target domain
Figure FDA0003568088060000012
And D1Common construction, i.e. Ds＝D1∪DaWherein n isaFor the number of samples in batch a, when a is 1, there is no classification prediction result of the previous batch due to the first model construction, so Ds＝D1The number of reconstructed source domain samples is:
Figure FDA0003568088060000013
the step 2) comprises the following steps:
step 21) inputting unlabeled a +1 batch target domain samples
Figure FDA0003568088060000014
ntIs the number of samples in the target domain;
step 22) analysis of X using principal ComponentssAnd XtDimension reduction to p dimension to generate subspace SsAnd St，Gd×pFor the set of all p-dimensional subspaces, each subspace is considered as the Grassman manifold space Gd×pAt one point above, let Φ (t) be Gd×pA geodesic line of (1), wherein t ∈ [0,1 ]]，Φ(0)＝SsAnd Φ (1) is StAs ends of a geodesic line, ziAnd zjIs xiAnd xjThe inner product of the projected feature vector in the infinite dimensional space is expressed as:
Figure FDA0003568088060000015
in the above formula xi,xj∈Ds∪DtAnd G denotes a geodesic core:
Figure FDA0003568088060000016
in the above formula RsFrom XsThe d-p dimensional feature composition, U, remaining after the P dimensional feature extraction by PCA1And U2Are orthogonal matrices and are solved by singular value decomposition1,Λ2And Λ3For a diagonal matrix, the element values in the matrix are respectively:
Figure FDA0003568088060000017
mixing XsAnd XtSample feature space usage after projection
Figure FDA0003568088060000018
And
Figure FDA0003568088060000019
is shown in which
Figure FDA00035680880600000110
Step 23) Using ZsTraining a classifier by a k nearest neighbor algorithm with k equal to 1, and then converting Z into ZtBrought into the classifier to obtain pseudo labels
Figure FDA0003568088060000021
Step 24) Using ZsAnd ZtSelecting Gaussian kernels to construct kernel functions
Figure FDA0003568088060000022
Step 25) determining Z by k-nearest neighbor algorithmsAnd ZtThe neighbor relation of each point in the similarity matrix W is obtained:
Figure FDA0003568088060000023
Figure FDA0003568088060000024
in the above formula r (z)i,zj) 1 represents ziAnd zjAre in a mutual neighbor relationship; after obtaining W, calculating a Laplace matrix L:
Figure FDA0003568088060000025
in the above formula, D is a diagonal matrix consisting of
Figure FDA0003568088060000026
And calculating to obtain, wherein the manifold regular constraint term is expressed as:
Figure FDA0003568088060000027
step 26) for ZsAnd ZtPerforming conditional distribution self-adaptation, measuring the conditional distribution difference on a regeneration Hilbert space through the maximum mean difference, and approximating statistical estimation by using an empirical estimation formula:
Figure FDA0003568088060000028
where C represents the total number of classes of labels contained within the sample,
Figure FDA0003568088060000029
consists of two diagonal matrices, wherein:
Figure FDA00035680880600000210
Figure FDA00035680880600000211
Figure FDA00035680880600000212
represents ZsAnd ZtN with the middle label being llSet of classification results for individual samples, f denotes Gd×pThe following classification prediction function:
Figure FDA0003568088060000031
in the above formula, alpha ═ alpha1,α2,...,αN]TFor coefficient vectors, the final calculation formula of the conditional distribution adaptive term is:
MMD2(HK,Qs,Qt)＝tr(fTMf)
in the above formula, each element in M is directly calculated by the following formula:
Figure FDA0003568088060000032
according to the structure risk minimization principle, combining a manifold regularization term and a condition distribution self-adaptive term, and finally optimizing a target of a classifier f as follows:
Figure FDA0003568088060000033
in the above equation, U is an indication matrix of the domain where the sample is located, that is:
Figure FDA0003568088060000034
let the partial derivative of α in the optimization equation be 0 to solve for α:
α＝(λ1I+(λ2M+λ3L+U)K)-1UYT
using f to complete pairs
Figure FDA0003568088060000035
Repeating step 26) e times to iteratively update M and α;
step 27) obtaining the prediction label of the sample of the batch
Figure FDA0003568088060000036
And storing the samples to wait for the input of the next batch of samples.
CN202010158546.5A 2020-03-09 2020-03-09 Online drift compensation method for sensor in bionic olfaction system Active CN111474297B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010158546.5A CN111474297B (en) 2020-03-09 2020-03-09 Online drift compensation method for sensor in bionic olfaction system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010158546.5A CN111474297B (en) 2020-03-09 2020-03-09 Online drift compensation method for sensor in bionic olfaction system

Publications (2)

Publication Number Publication Date
CN111474297A CN111474297A (en) 2020-07-31
CN111474297B true CN111474297B (en) 2022-05-03

Family

ID=71747245

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010158546.5A Active CN111474297B (en) 2020-03-09 2020-03-09 Online drift compensation method for sensor in bionic olfaction system

Country Status (1)

Country Link
CN (1) CN111474297B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113030197B (en) * 2021-03-26 2022-11-04 哈尔滨工业大学 Gas sensor drift compensation method

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5177994A (en) * 1991-05-22 1993-01-12 Suntory Limited And Tokyo Institute Of Technology Odor sensing system
CN1514239A (en) * 2003-07-23 2004-07-21 西安交通大学 Method of gas sensor drift inhibition
CN102507676A (en) * 2011-11-01 2012-06-20 重庆大学 On-line drift compensation method of electronic nose based on multiple self-organizing neural networks
CN103544392A (en) * 2013-10-23 2014-01-29 电子科技大学 Deep learning based medical gas identifying method
WO2015127085A1 (en) * 2014-02-19 2015-08-27 Ino Therapeutics Llc Methods for compensating long term sensitivity drift of electrochemical gas sensors exposed to nitric oxide
CN105891422A (en) * 2016-04-08 2016-08-24 重庆大学 Electronic nose gas identification method based on source domain migration extreme learning to realize drift compensation
CN109472303A (en) * 2018-10-30 2019-03-15 浙江工商大学 A kind of gas sensor drift compensation method based on autoencoder network decision
CN110146655A (en) * 2019-05-31 2019-08-20 重庆大学 A kind of anti-electronic nose bleach-out process based on adaptive sub-space learning normal form
CN110222785A (en) * 2019-06-13 2019-09-10 重庆大学 Self-adapting confidence degree Active Learning Method for gas sensor drift correction

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5177994A (en) * 1991-05-22 1993-01-12 Suntory Limited And Tokyo Institute Of Technology Odor sensing system
CN1514239A (en) * 2003-07-23 2004-07-21 西安交通大学 Method of gas sensor drift inhibition
CN102507676A (en) * 2011-11-01 2012-06-20 重庆大学 On-line drift compensation method of electronic nose based on multiple self-organizing neural networks
CN103544392A (en) * 2013-10-23 2014-01-29 电子科技大学 Deep learning based medical gas identifying method
WO2015127085A1 (en) * 2014-02-19 2015-08-27 Ino Therapeutics Llc Methods for compensating long term sensitivity drift of electrochemical gas sensors exposed to nitric oxide
CN105891422A (en) * 2016-04-08 2016-08-24 重庆大学 Electronic nose gas identification method based on source domain migration extreme learning to realize drift compensation
CN109472303A (en) * 2018-10-30 2019-03-15 浙江工商大学 A kind of gas sensor drift compensation method based on autoencoder network decision
CN110146655A (en) * 2019-05-31 2019-08-20 重庆大学 A kind of anti-electronic nose bleach-out process based on adaptive sub-space learning normal form
CN110222785A (en) * 2019-06-13 2019-09-10 重庆大学 Self-adapting confidence degree Active Learning Method for gas sensor drift correction

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Domain Adaptation Extreme Learning Machines for Drift Compensation in E-Nose Systems;Zhang Lei等;《IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT》;20150731;第64卷(第7期);1790-1801 *
基于域校正的电子鼻传感器漂移抑制方法研究;徐娟;《中国优秀博硕士学问论文全文数据库（硕士） 信息科技辑》;20200215(第6期);I140-343 *
基于堆栈自编码网络的气体传感器漂移补偿方法;宋艳玲;《中国优秀博硕士学位论文全文数据库（硕士） 信息科技辑》;20190615(第6期);I140-139 *

Also Published As

Publication number Publication date
CN111474297A (en) 2020-07-31

Similar Documents

Publication Publication Date Title
CN109492830B (en) Mobile pollution source emission concentration prediction method based on time-space deep learning
Labach et al. Survey of dropout methods for deep neural networks
CN111999649A (en) XGboost algorithm-based lithium battery residual life prediction method
CN110675632B (en) Vehicle short-time trajectory prediction control method aiming at multi-feature space and data sparseness
CN114676742A (en) Power grid abnormal electricity utilization detection method based on attention mechanism and residual error network
CN112763967B (en) BiGRU-based intelligent electric meter metering module fault prediction and diagnosis method
CN112949828A (en) Graph convolution neural network traffic prediction method and system based on graph learning
CN111369045A (en) Method for predicting short-term photovoltaic power generation power
CN111474297B (en) Online drift compensation method for sensor in bionic olfaction system
CN114019370A (en) Motor fault detection method based on gray level image and lightweight CNN-SVM model
CN114863151B (en) Image dimension reduction clustering method based on fuzzy theory
CN116168548A (en) Traffic flow prediction method of space-time attention pattern convolution network based on multi-feature fusion
CN114781723A (en) Short-term photovoltaic output prediction method based on multi-model fusion
CN112700432A (en) Texture surface defect detection method and system based on abnormal synthesis and decomposition
CN114299324B (en) Pathological image classification method and system based on multiscale domain countermeasure network
CN115294550A (en) Automatic driving automobile road scene understanding method based on multi-task learning
CN114596726B (en) Parking berth prediction method based on interpretable space-time attention mechanism
CN117436595B (en) New energy automobile energy consumption prediction method, device, equipment and storage medium
CN114580262A (en) Lithium ion battery health state estimation method
CN112927507A (en) Traffic flow prediction method based on LSTM-Attention
CN116484275A (en) Aluminum electrolysis anode effect monitoring method based on self-adaptive space-time diagram convolutional neural network
CN115759343A (en) E-LSTM-based user electric quantity prediction method and device
CN113469013A (en) Motor fault prediction method and system based on transfer learning and time sequence
Zhu et al. Residual attention dual autoencoder for anomaly detection and localization in cigarette packaging
CN115131588B (en) Image robust clustering method based on fuzzy clustering

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