CN110794090A - Emotion electronic nose implementation method - Google Patents
Emotion electronic nose implementation method Download PDFInfo
- Publication number
- CN110794090A CN110794090A CN201911006190.7A CN201911006190A CN110794090A CN 110794090 A CN110794090 A CN 110794090A CN 201911006190 A CN201911006190 A CN 201911006190A CN 110794090 A CN110794090 A CN 110794090A
- Authority
- CN
- China
- Prior art keywords
- emotion
- smell
- odor
- fuzzy
- electronic nose
- 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.)
- Pending
Links
- 230000008451 emotion Effects 0.000 title claims abstract description 107
- 238000000034 method Methods 0.000 title claims abstract description 32
- 235000019645 odor Nutrition 0.000 claims abstract description 75
- 239000013598 vector Substances 0.000 claims abstract description 27
- 239000000463 material Substances 0.000 claims abstract description 18
- 230000004044 response Effects 0.000 claims abstract description 16
- 230000001537 neural effect Effects 0.000 claims abstract description 12
- 230000007436 olfactory function Effects 0.000 claims abstract description 5
- 230000006870 function Effects 0.000 claims description 18
- 238000005070 sampling Methods 0.000 claims description 11
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 238000007781 pre-processing Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 4
- 238000011084 recovery Methods 0.000 claims description 4
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 230000008909 emotion recognition Effects 0.000 claims description 2
- 238000004590 computer program Methods 0.000 claims 1
- 230000035943 smell Effects 0.000 description 44
- 239000007789 gas Substances 0.000 description 17
- 230000002996 emotional effect Effects 0.000 description 14
- 238000011160 research Methods 0.000 description 10
- 238000001514 detection method Methods 0.000 description 7
- 238000012360 testing method Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 3
- 238000001914 filtration Methods 0.000 description 3
- 230000003993 interaction Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 229910044991 metal oxide Inorganic materials 0.000 description 2
- 150000004706 metal oxides Chemical class 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000003909 pattern recognition Methods 0.000 description 2
- 239000002304 perfume Substances 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 239000000047 product Substances 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000008786 sensory perception of smell Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 244000144730 Amygdalus persica Species 0.000 description 1
- 101100054292 Arabidopsis thaliana ABCG36 gene Proteins 0.000 description 1
- 101100351526 Arabidopsis thaliana PEN3 gene Proteins 0.000 description 1
- 206010063659 Aversion Diseases 0.000 description 1
- 241000628997 Flos Species 0.000 description 1
- 241000208152 Geranium Species 0.000 description 1
- 241000282412 Homo Species 0.000 description 1
- 244000178870 Lavandula angustifolia Species 0.000 description 1
- 235000010663 Lavandula angustifolia Nutrition 0.000 description 1
- 241000234479 Narcissus Species 0.000 description 1
- 206010028813 Nausea Diseases 0.000 description 1
- 235000006040 Prunus persica var persica Nutrition 0.000 description 1
- 241000220317 Rosa Species 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 235000013361 beverage Nutrition 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 210000003608 fece Anatomy 0.000 description 1
- 235000013305 food Nutrition 0.000 description 1
- 239000001102 lavandula vera Substances 0.000 description 1
- 235000018219 lavender Nutrition 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000008693 nausea Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 238000006722 reduction reaction Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000035807 sensation Effects 0.000 description 1
- 235000019615 sensations Nutrition 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000037152 sensory function Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000000638 stimulation Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 210000004243 sweat Anatomy 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 239000000341 volatile oil Substances 0.000 description 1
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
- G01N1/00—Sampling; Preparing specimens for investigation
- G01N1/02—Devices for withdrawing samples
- G01N1/22—Devices for withdrawing samples in the gaseous state
-
- 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
- G01N33/0009—General constructional details of gas analysers, e.g. portable test equipment
- G01N33/0027—General constructional details of gas analysers, e.g. portable test equipment concerning the detector
- G01N33/0031—General constructional details of gas analysers, e.g. portable test equipment concerning the detector comprising two or more sensors, e.g. a sensor array
-
- 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
- G01N33/0009—General constructional details of gas analysers, e.g. portable test equipment
- G01N33/0062—General constructional details of gas analysers, e.g. portable test equipment concerning the measuring method or the display, e.g. intermittent measurement or digital display
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/043—Architecture, e.g. interconnection topology based on fuzzy logic, fuzzy membership or fuzzy inference, e.g. adaptive neuro-fuzzy inference systems [ANFIS]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Pathology (AREA)
- Immunology (AREA)
- Biochemistry (AREA)
- Analytical Chemistry (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Food Science & Technology (AREA)
- Biomedical Technology (AREA)
- Medicinal Chemistry (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Combustion & Propulsion (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Automation & Control Theory (AREA)
- Computational Mathematics (AREA)
- Fuzzy Systems (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to an emotion electronic nose implementation method and electronic equipment, which comprise the following steps: different odors were collected and a library of odor materials was constructed. And selecting healthy people with normal olfactory function to smell each smell in the smell material library respectively, and scoring each basic emotion according to the emotion condition induced by the smell. And (3) calculating the average score of each basic emotion of each smell according to the scoring values of all the subjects on the smell, and normalizing to obtain the emotion distribution vector which can be induced by the smell and is used as the emotion label of the smell. Each smell in the smell material library is sampled by a sensor array in the electronic nose system, and a corresponding sensor array response curve is obtained. An odor feature database with emotion labels is constructed. And combining a fuzzy clustering method with a self-adaptive neural fuzzy reasoning system to construct a smell emotion model capable of predicting emotion distribution vectors.
Description
Technical Field
The invention belongs to the cross field of instrument measurement and emotion calculation, and relates to an electronic nose implementation method with an emotion function.
Background
Olfaction is the sensation caused by the stimulation of nose receptors by chemical molecules and is the oldest sensory function in the history of biological evolution. Inspired by the biological olfactory principle, researchers in the eighties of the twentieth century have proposed the concept of a gas detection device that mimics the mammalian olfactory system in structure and function, called an electronic nose (Gardner J w. pattern recognition in the water electronic nose, University of water, UK, 1988). The device mainly comprises a gas sensor array, an information preprocessing module and a pattern recognition system, and can realize the detection and recognition of simple or complex odor. The electronic nose has been widely researched and applied in the industries of environmental protection, petroleum/chemical industry, food/beverage, perfume, medical treatment and the like because of the advantages of portability, high detection speed, short response time and the like. For example, a malodorous electronic nose instrument and a malodorous gas multi-point centralized online monitoring method, patent numbers CN108709955A,2018, have developed a malodorous electronic nose instrument which can monitor malodorous gas online. A detachable portable electronic nose system for on-site rapid detection of agricultural products was developed by a wang jun professor team of the university of zhejiang (wang jun, xukeming, detachable portable electronic nose for on-site rapid detection of agricultural products, patent numbers CN107045005A, 2017). The method is characterized in that a professor team of luodehan in luodehan of the university of Guangdong industry (Fan Dan Jun, Luodehan, Yuhao) researches a classification and identification method of pungent traditional Chinese medicinal materials based on an electronic nose, the institute of Guangdong industry university, 2015,32(03):91-96) uses a PEN3 electronic nose to classify the odor of the pungent traditional Chinese medicinal materials, and mode classification of volatile odor information of different types of pungent traditional Chinese medicinal materials is achieved. A Mengqinghao professor team of Tianjin university (Mengqinghao, Li Zhihua, Qi cultivate the front, Zengming, Zhao is. a hand-held electronic nose for on-line liquor identification, patent No. CN10742266987. 2017) continuously optimizes and improves from the aspects of sensor array, gas circuit/gas chamber, hardware circuit, sampling method, system volume and the like aiming at the portable requirement of liquor detection application, successively develops a desktop split type, portable, hand-held and Mini type four-generation electronic nose system, and realizes the identification of the liquor type.
The current electronic nose system is generally mainly used for detection and recognition of smells, and cannot induce different emotions by smelling different smells like a human. Scientific studies have shown that the sense of smell is natural and emotional, that the smell stimuli can induce different emotional states, such as narcissus and violet scents that make people warm and pleasant, that geranium scents can relieve people from tension, and that the musty smell makes people dislike, uneasy, etc. (Wen, Von fruit, olfactory perception and their interaction with the emotional system. psychology advances, 2012,20(1): 2-9.). However, unlike the human nose, most of the existing electronic nose systems do not have the same emotional ability as a human, and do not truly realize the imitation of the biological olfactory system.
The Rafi Haddad doctor of the department of neurobiology, wilm, research (Rafi Haddad, AbebeMedhanie, Yehudah Roth, et al, predicting Odor great research with an electronic nose, plos computerized Biology,2010,6(4): e1000740) at first proposed that the electronic nose was no longer trained to recognize specific odors, but was trained from an Odor pleasure perspective to estimate the pleasure of the odors. The team of the doctor Rafi Haddad led to the development of an electronic nose system that predicts the pleasantness of strange smells like a human, i.e. the electronic nose can predict whether the smell to be measured smells pleasant or unpleasant, or at any point between these two endpoints. At present, no electronic nose system research report for combining the electronic nose with odor-induced emotion to enable the electronic nose to have the emotional ability like a human is available in China.
Disclosure of Invention
The invention aims to provide a realization method for enabling an electronic nose to have emotional ability like a human, so that the electronic nose can predict which emotions can be induced by the odor to be detected, recognize main emotions and approach a biological olfactory system better. The technical scheme is as follows:
an implementation method of an emotion electronic nose comprises the following steps:
1) different odors were collected and a library of odor materials was constructed.
2) And selecting healthy people with normal olfactory function to smell each smell in the smell material library respectively, and scoring each basic emotion according to the emotion condition induced by the smell.
3) And (3) calculating the average score of each basic emotion of each smell according to the scoring values of all the subjects on the smell, and normalizing to obtain the emotion distribution vector which can be induced by the smell and is used as the emotion label of the smell.
4) Each smell in the smell material library is sampled by a sensor array in the electronic nose system, and a corresponding sensor array response curve is obtained.
5) Preprocessing the original data of the odor sample acquired in the step 4), extracting a plurality of characteristic values from each preprocessed odor sensor response curve to serve as characteristic vectors of the odor, and constructing an odor characteristic database with emotion labels based on the extracted characteristics and the emotion labels corresponding to the step 3).
6) And combining a fuzzy clustering method with a self-adaptive neural fuzzy reasoning system to construct a smell emotion model capable of predicting emotion distribution vectors.
7) Inputting a part of the odor characteristic data obtained in the step 5) into the odor emotion model constructed in the step 6) as a training set, training the odor emotion model, taking the rest of the characteristic data as a test set, testing the trained odor emotion model and calculating the accuracy. And 7) repeating the step 7) for multiple times, and calculating the average identification accuracy of the constructed scent emotion model.
8) And performing emotion recognition on the odor to be detected by using the trained odor emotion model, wherein the output result of the electronic nose is an emotion distribution vector which can be induced by the odor to be detected, and the highest component value is the main emotion.
Preferably, the extracted characteristic values include a steady-state response value, a peak value, a mean value, a recovery time, a steady-state response time, a difference value, a first-order differential maximum value, a value corresponding to the first-order differential maximum value, a first-order differential minimum value, and a second-order differential maximum value; taking the obtained 100-dimensional characteristic value as a characteristic vector of the smell;
the adaptive neural fuzzy inference system adopts a first-order Takagi-Sugeno-Kang (TSK) fuzzy system based on data driving.
The construction steps of the smell emotion model are as follows: (a) firstly, determining the number of fuzzy rules and the membership function type used by the characteristics; (b) selecting the input features of each rule by using a fuzzy subspace algorithm, and determining membership function parameters of each feature used under each rule; (c) fuzzifying the screened feature vectors, and inputting the fuzzified feature vectors into a self-adaptive neural fuzzy inference system; (d) the self-adaptive fuzzy inference system carries out fuzzy inference according to the established fuzzy rule and the input fuzzy set to obtain an inference result; (e) and performing defuzzification and normalization on the inference result to obtain a predicted emotion distribution vector.
The invention has the following beneficial effects:
1) functionally, the implementation method for enabling the electronic nose to have the human-like emotional ability is characterized in that the electronic nose can predict emotions which can be induced by the smell to be detected, can give the proportion of each emotion and identify main emotion, and is a supplement to the functions of the existing electronic nose system, so that the electronic nose is closer to a biological olfactory system.
2) In the aspect of algorithm design, in consideration of fuzzy complexity of emotion, the scent emotion model is constructed on the basis of a fuzzy clustering algorithm and a self-adaptive neural fuzzy reasoning system. Particularly, the input features of each rule are selected by using a fuzzy clustering algorithm, so that the noise features which have adverse effects on modeling can be removed, the dimensionality of the input features is reduced, each rule can be inferred from different visual angles, and the interpretability of a fuzzy system is enhanced. In addition, the fuzzy clustering algorithm can also be used for determining membership function parameters of the input features; by adopting the data-driven type self-adaptive neural fuzzy inference system, knowledge can be acquired from data to construct a fuzzy inference rule.
In summary, the electronic nose system provided by the invention has important potential application values, such as 1) in the research aspect of intelligent robots, the electronic nose with emotion function is carried on the robot system, so that the robot has similar olfactory emotion function to human, and the development of robot intelligence is promoted; 2) provides a new method for rapid smell screening and environmental monitoring in the perfume industry.
Drawings
Fig. 1 is a block diagram of an electronic nose structure adopted by the present invention.
FIG. 2 is a flowchart of an implementation method of the emotion electronic nose provided by the present invention.
FIG. 3 is a diagram of a scent emotion model according to the present invention.
Detailed Description
The current electronic nose system is mainly used for detecting and identifying smell, does not have the function of generating different emotions by smelling different smells like a human, and does not realize the simulation of a biological olfactory system in a real sense. In order to improve the functions of the existing electronic nose system, the patent provides an implementation method for enabling the electronic nose to have emotional ability. Particularly, in consideration of fuzzy complex characteristics of emotions, the method constructs an odor emotion model based on fuzzy clustering and a self-adaptive neural fuzzy inference system, and programs a model algorithm into the electronic nose, so that the electronic nose can predict which emotions can be induced by the odor to be detected, can give the proportion of each emotion, identifies main emotions, and is closer to a biological olfactory system. The implementation method for enabling the electronic nose to have the emotional function mainly comprises the following steps: 1) collecting a plurality of gases with different odors and establishing an odor material library; 2) selecting healthy people with normal olfactory function to score basic emotions which can be induced by each odor, calculating the average score of each basic emotion under the odor, normalizing the average score to obtain an odor emotion distribution vector, and using the odor emotion distribution vector as an emotion label of the odor; 3) acquiring a signal of the odor in the odor material library by using an electronic nose system, and carrying out filtering, noise reduction and normalization pretreatment on the odor; 4) extracting features from the preprocessed smell data, and constructing a smell feature database with smell emotion labels by combining the smell emotion labels in the step 2); 5) constructing a smell emotion model based on fuzzy clustering and a self-adaptive neural fuzzy inference system; 6) training the odor emotion model constructed in the step 5) by taking part of the data in the step 4) as a training set, and using the rest data as a test set for testing the accuracy of the model; 7) and inputting the odor to be detected into the trained odor emotion model, and predicting the emotion label of the odor, wherein the emotion corresponding to the maximum value in the label is the main emotion.
The present invention will be described in detail with reference to the following embodiments and accompanying drawings. The embodiments are specific implementations on the premise of the technical scheme of the invention, and detailed implementation modes and processes are given. The scope of protection of the claims of the present application is not limited by the description of the embodiments below.
The structure of the electronic nose system adopted by the invention is shown in figure 1. The electronic nose system mainly comprises three parts, namely a sampling device, a sensor air chamber reaction device and a control and data acquisition preprocessing system. The work flow of the electronic nose is as follows: firstly, a dynamic headspace sampling method is utilized, a speed-adjustable miniature vacuum air pump brings headspace gas to be detected of a sampling bottle into a sensor air chamber along a sampling air path through filtered clean air, the sample gas to be detected reacts with a sensor array in the sensor air chamber, the sensor array collects odor information, odor signals are converted into electric signals, an AD chip stores and uploads array output voltage signals after sampling is finished, and an ARM processor carries out preprocessing and mode recognition on sampling data to obtain a recognition result through final analysis. The LCD touch screen is used for displaying information and receiving commands, and after the LCD touch screen is started to calibrate the screen, the operation of a user can be responded according to the programmed human-computer interaction interface, the state information of the electronic nose can be displayed, online drawing and online identification can be carried out, the result can be displayed, and the like.
The work flow diagram of the present invention is shown in fig. 2. The implementation method for enabling the electronic nose to have the emotional function is explained in detail below, and mainly comprises the following steps:
1) a library of scented materials is established. Selecting 13 odors, wherein 5 odors are T & T odor solution (rose odor, roasted coffee odor, sweat odor, peach odor and feces odor), and the rest 8 odors are herba Menthae, folium Camelliae sinensis, coffee, herba Rosmarini officinalis, flos Jasmini sambac, fructus Citri Limoniae, herba Vanillae and Lavender essential oil.
2) A scent sentiment label is obtained. 100 healthy persons with normal olfactory function (male and female halves) were selected and college students sniffed each smell in the smell material library separately and scored each basic emotion (happy, angry, surprise, nausea, fear, sadness) on a graduated scale from 0 to 1 according to the emotional condition induced by each smell. Wherein a score of 0 indicates the lowest degree of correlation and a score of 1 indicates the highest degree of correlation. And collecting scoring information of each smell, calculating an average value, normalizing to obtain an emotion distribution vector corresponding to the smell, and taking the emotion distribution vector as an emotion label of the smell.
3) And collecting odor signals. The sensor array of the electronic nose system adopted by the invention is composed of 10 metal oxide semiconductor gas sensors, 13 odors in the odor library are sampled, each odor is tested for 50 times, and 650 sample data are obtained by the 13 odors.
4) An odor feature database with emotion labels is constructed. Firstly, respectively carrying out noise reduction, filtering and normalization preprocessing operations on 650 odor sample data acquired in the step 3); next, for each scent, the following 10 features were extracted from the 10 pre-processed sensor response curves in the sensor array, respectively: the method comprises the following steps of (1) obtaining a steady-state response value, a peak value, a mean value, recovery time, steady-state response time, a difference value, a first-order differential maximum value, a value corresponding to the first-order differential maximum value, a first-order differential minimum value and a second-order differential maximum value; taking the obtained 100-dimensional characteristic value as a characteristic vector of the smell; and finally, constructing an odor characteristic database with emotion labels based on the extracted characteristics and the emotion labels corresponding to the step 2).
5) And constructing a smell emotion model based on fuzzy clustering and a self-adaptive neural fuzzy inference system. The fuzzy clustering algorithm adopts a fuzzy subspace clustering method which can convert a high-dimensional data space into a related subspace; the adaptive neural fuzzy inference system adopts a first-order Takagi-Sugeno-Kang (TSK) fuzzy system based on data driving. The IF-THEN fuzzy rule of the first-order TSK fuzzy system, the K-th fuzzy rule is expressed as follows:
wherein, IF represents before the ruleWhere, THEN denotes the rule back, K is the fuzzy rule number, x ═ x1,x2,x3,…,xd]Is the d-dimensional feature vector of the input,
is represented by the formulaiThe fuzzy subset in the corresponding k-th rule, Λ is the conjunction operator, fk(x) Is the output value of the k-th rule,is a parameter of the rule back-piece. The odor emotion model is shown in fig. 3 and comprises the following two main parts: (a) and learning the rule antecedents by using a fuzzy subspace clustering algorithm. Firstly, setting the number of the cluster categories as 6 and the number of the fuzzy rules as 6 according to the basic emotion, and selecting the membership function as a Gaussian membership function. Secondly, aiming at each rule, important features are selected from the input feature vectors by using a fuzzy subspace clustering method. For example, after being filtered, the feature vector corresponding to the input of the k rule becomes
And determining the Gaussian membership function parameters of the characteristics used under each rule according to the clustering result. (b) And (3) carrying out rule postware learning by using a Levenberg-Marquart (LM) algorithm, and optimizing rule postware parameters.
6) And randomly selecting 80% of samples from the smell characteristic database in the step 4) as a training set, training the smell emotion model established in the step 5), inputting the rest 20% of samples as a test set into the trained model, and calculating the recognition accuracy. This process was repeated 10 times, and the results were averaged 10 times to determine the recognition accuracy of the odor model.
7) Given a test odor data, the following 10 features were extracted: the method comprises the following steps of (1) obtaining a steady-state response value, a peak value, a mean value, recovery time, steady-state response time, a difference value, a first-order differential maximum value, a value corresponding to the first-order differential maximum value, a first-order differential minimum value and a second-order differential maximum value; inputting the obtained 100-dimensional characteristic value as the characteristic vector of the smell into the trained model in the step 6), and displaying the result through an LCD.
The electronic nose system device adopted by the invention comprises an air pump, a sampling bottle, a sensor air chamber, a sensor array, an electromagnetic valve, a signal preprocessing module, an embedded processing module and an LCD touch screen. The device is characterized in that the air pump brings the headspace gas of the sampling bottle into the sensor air chamber along the sampling gas path through the filtered clean air; the sensor gas chamber is used for collecting gas in the gas chamber and converting a gas signal into a corresponding electric signal; the sensor array has higher response amplitude to the odor of the odor material library, and comprises a plurality of sensors, such as metal oxide sensors with cross sensitivity, temperature and humidity sensors and the like, and is used for collecting different volatile gases; the signal preprocessing module is used for filtering, denoising and analog-to-digital converting the electric signals generated by the sensor array; the embedded processing module is used for carrying out feature extraction, mode classification and storage on output signals of the signal preprocessing module, and the embedded processing module is loaded with a smell emotion model. The LCD touch screen provides a human-computer interaction interface for receiving commands and displaying information.
The main difference between the research of the patent and the research of the Rafi Haddad doctor lies in that: 1) human emotion is rich and complex, such as half-joy and sorrow, and the like. The electronic nose is trained only from the perspective of smell pleasure by the doctor Rafi Haddad, and although the developed electronic nose system can classify and score the pleasure of the smell to be detected, a certain distance exists between the electronic nose system and the electronic nose system for achieving the goal that the electronic nose has the human-like emotional ability. In theory of emotion, psychologist Ekman (Ekman P, friend WV. Unmasking the face: a guide to recogniting emotions from faces, 1975, Prentice-Hall, Oxford) presented a basic theory of emotion in 1975. Ekman states that humans have six basic emotions (happy, angry, sad, aversion, fear, surprise), and other emotions are formed by mixing two or more of the six basic emotions. Since then, in the field of emotional research, numerous scholars have developed research based on the basic theory of emotion of Ekman. Unlike the research of doctor RafiHaddad from the perspective of smell pleasure, the patent makes the electronic nose have the human-like emotional function as much as possible based on the basic human emotion theory proposed by Ekman. 2) Considering that human emotions have fuzzy complexity, the scent emotion model is constructed based on the fuzzy theory, so that the electronic nose can predict which emotions can be induced by the scent to be detected, can give the proportion of each emotion and can identify main emotions.
Claims (5)
1. An implementation method of an emotion electronic nose comprises the following steps:
1) different odors were collected and a library of odor materials was constructed.
2) Selecting healthy people with normal olfactory function to smell each smell in the smell material library respectively, and scoring each basic emotion according to the emotion condition induced by the smell;
3) for each smell, calculating each basic emotion average score of the smell according to the scoring values of all the subjects on the smell, normalizing to obtain an emotion distribution vector which can be induced by the smell and using the emotion distribution vector as an emotion label;
4) sampling each smell in the smell material library through a sensor array in the electronic nose system to obtain a sensor array response curve corresponding to the smell material library;
5) preprocessing the original data of the odor sample acquired in the step 4), extracting a plurality of characteristic values from each preprocessed odor sensor response curve to serve as characteristic vectors of the odor, and constructing an odor characteristic database with emotion labels based on the extracted characteristics and the emotion labels corresponding to the step 3);
6) combining a fuzzy clustering method with a self-adaptive neural fuzzy inference system to construct a smell emotion model capable of predicting emotion distribution vectors;
7) training a smell emotion model by using the smell characteristic data obtained in the step 5);
8) and performing emotion recognition on the odor to be detected by using the trained odor emotion model, wherein the output result of the electronic nose is an emotion distribution vector which can be induced by the odor to be detected, and the highest component value is the main emotion.
2. The method of claim 1, wherein the extracted feature values include a steady-state response value, a peak value, a mean value, a recovery time, a steady-state response time, a difference value, a first-order differential maximum value, a value corresponding to a first-order differential maximum value, a first-order differential minimum value, and a second-order differential maximum value.
3. The method according to claim 1, wherein the adaptive neuro-fuzzy inference system employs a first order Takagi-Sugeno-kang (tsk) fuzzy system based on data driving.
4. The method of claim 1, wherein the odor emotion model is constructed by the steps of: (a) firstly, determining the number of fuzzy rules and the membership function type used by the characteristics; (b) selecting the input features of each rule by using a fuzzy subspace algorithm, and determining membership function parameters of each feature used under each rule; (c) fuzzifying the screened feature vectors, and inputting the fuzzified feature vectors into a self-adaptive neural fuzzy inference system; (d) the self-adaptive fuzzy inference system carries out fuzzy inference according to the established fuzzy rule and the input fuzzy set to obtain an inference result; (e) and performing defuzzification and normalization on the inference result to obtain a predicted emotion distribution vector.
5. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the method of any of claims 1-4 are implemented when the program is executed by the processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911006190.7A CN110794090A (en) | 2019-10-22 | 2019-10-22 | Emotion electronic nose implementation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911006190.7A CN110794090A (en) | 2019-10-22 | 2019-10-22 | Emotion electronic nose implementation method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110794090A true CN110794090A (en) | 2020-02-14 |
Family
ID=69440544
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911006190.7A Pending CN110794090A (en) | 2019-10-22 | 2019-10-22 | Emotion electronic nose implementation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110794090A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111426801A (en) * | 2020-05-09 | 2020-07-17 | 上海宁和环境科技发展有限公司 | Electronic nose learning and domesticating method and equipment thereof |
| CN112036482A (en) * | 2020-08-31 | 2020-12-04 | 重庆大学 | Traditional Chinese medicine classification method based on electronic nose sensor data |
| CN113420443A (en) * | 2021-06-23 | 2021-09-21 | 天津市生态环境科学研究院(天津市环境规划院、天津市低碳发展研究中心) | Accurate stink simulation method coupled with peak-to-average factor |
| CN114527198A (en) * | 2020-10-30 | 2022-05-24 | 中国石油化工股份有限公司 | Polymer product odor detection method and device and electronic nose system |
| CN115643087A (en) * | 2022-10-24 | 2023-01-24 | 天津大学 | DNS tunnel detection method based on fusion of coding characteristics and statistical behavior characteristics |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011021046A1 (en) * | 2009-08-19 | 2011-02-24 | University Of Leicester | Fuzzy inference methods and apparatus, systems and apparatuses using such inference apparatus |
| CN103630698A (en) * | 2013-12-03 | 2014-03-12 | 杭州协正信息技术有限公司 | Stereoscopic electronic nose for simulating animal olfactory organ structure |
| CN106596860A (en) * | 2016-12-19 | 2017-04-26 | 深圳市北测检测技术有限公司 | Detection method and detection system for automobile smell |
| US20170343521A1 (en) * | 2016-05-26 | 2017-11-30 | Electronics And Telecommunications Research Institute | Apparatus and method for generating olfactory information |
| CN110333319A (en) * | 2019-06-28 | 2019-10-15 | 天津大学 | The interior ppb grades of low concentration oder levels evaluation methods based on hand-hold electric nasus |
| CN112927763A (en) * | 2021-03-05 | 2021-06-08 | 广东工业大学 | Prediction method for odor descriptor rating based on electronic nose |
-
2019
- 2019-10-22 CN CN201911006190.7A patent/CN110794090A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011021046A1 (en) * | 2009-08-19 | 2011-02-24 | University Of Leicester | Fuzzy inference methods and apparatus, systems and apparatuses using such inference apparatus |
| CN103630698A (en) * | 2013-12-03 | 2014-03-12 | 杭州协正信息技术有限公司 | Stereoscopic electronic nose for simulating animal olfactory organ structure |
| US20170343521A1 (en) * | 2016-05-26 | 2017-11-30 | Electronics And Telecommunications Research Institute | Apparatus and method for generating olfactory information |
| CN106596860A (en) * | 2016-12-19 | 2017-04-26 | 深圳市北测检测技术有限公司 | Detection method and detection system for automobile smell |
| CN110333319A (en) * | 2019-06-28 | 2019-10-15 | 天津大学 | The interior ppb grades of low concentration oder levels evaluation methods based on hand-hold electric nasus |
| CN112927763A (en) * | 2021-03-05 | 2021-06-08 | 广东工业大学 | Prediction method for odor descriptor rating based on electronic nose |
Non-Patent Citations (4)
| Title |
|---|
| RAFI HADDAD等: "Predicting Odor Pleasantness with an Electronic Nose", 《PLOS COMPUTATIONAL BIOLOGY》 * |
| 刘士荣和俞金寿: "基于最优模糊聚类的模糊推理系统及其基于最优模糊聚类的模糊推理系统及其基于最优模糊聚类的模糊推理系统及其在产品质量估计中的应用", 《信息与控制》 * |
| 曹勤: "神经网络在人工嗅觉信息处理技术中的应用研究", 《中国优秀硕士学位论文全文数据库》 * |
| 爱德华·E.史密斯，斯蒂芬·M.科斯林: "《认知心理学 心智与脑》", 31 July 2017 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111426801A (en) * | 2020-05-09 | 2020-07-17 | 上海宁和环境科技发展有限公司 | Electronic nose learning and domesticating method and equipment thereof |
| CN111426801B (en) * | 2020-05-09 | 2022-08-02 | 上海宁和环境科技发展有限公司 | Electronic nose learning and domesticating method and equipment thereof |
| CN112036482A (en) * | 2020-08-31 | 2020-12-04 | 重庆大学 | Traditional Chinese medicine classification method based on electronic nose sensor data |
| CN112036482B (en) * | 2020-08-31 | 2023-10-24 | 重庆大学 | Traditional Chinese medicine classification method based on electronic nose sensor data |
| CN114527198A (en) * | 2020-10-30 | 2022-05-24 | 中国石油化工股份有限公司 | Polymer product odor detection method and device and electronic nose system |
| CN113420443A (en) * | 2021-06-23 | 2021-09-21 | 天津市生态环境科学研究院(天津市环境规划院、天津市低碳发展研究中心) | Accurate stink simulation method coupled with peak-to-average factor |
| CN113420443B (en) * | 2021-06-23 | 2022-03-01 | 天津市生态环境科学研究院(天津市环境规划院、天津市低碳发展研究中心) | Accurate stink simulation method coupled with peak-to-average factor |
| CN115643087A (en) * | 2022-10-24 | 2023-01-24 | 天津大学 | DNS tunnel detection method based on fusion of coding characteristics and statistical behavior characteristics |
| CN115643087B (en) * | 2022-10-24 | 2024-04-30 | 天津大学 | DNS tunnel detection method based on fusion of coding features and statistical behavior features |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110794090A (en) | Emotion electronic nose implementation method | |
| Chen et al. | Gas recognition in E-nose system: A review | |
| Chang et al. | ScalpEye: A deep learning-based scalp hair inspection and diagnosis system for scalp health | |
| Chen et al. | Discrimination of green tea quality using the electronic nose technique and the human panel test, comparison of linear and nonlinear classification tools | |
| Jing et al. | A bioinspired neural network for data processing in an electronic nose | |
| LAWLESS et al. | Consistency of multidimensional scaling models derived from odor sorting | |
| Alonso et al. | Cooperation between expert knowledge and data mining discovered knowledge: Lessons learned | |
| CN112185493A (en) | Personality preference diagnosis device and project recommendation system based on same | |
| CN115251928A (en) | University student mental health assessment system based on deep learning | |
| Xiong et al. | Image presentation and effective classification of odor intensity levels using multi-channel electronic nose technology combined with GASF and CNN | |
| Purbawa et al. | Adaptive filter for detection outlier data on electronic nose signal | |
| Jing et al. | Data-driven implicit design preference prediction model for product concept evaluation via BP neural network and EEG | |
| CN117995399B (en) | Large-scale mental health intelligent evaluation method and system based on multi-modal data | |
| Perangin-Angin et al. | Classification of stress in office work activities using extreme learning machine algorithm and one-way anova f-test feature selection | |
| Hou et al. | A triangular hashing learning approach for olfactory EEG signal recognition | |
| Al-Mulla et al. | Classification of localized muscle fatigue with genetic programming on sEMG during isometric contraction | |
| Aziz et al. | Classification of Human Concentration in EEG Signals using Hilbert Huang Transform. | |
| Searle et al. | System identification of electronic nose data from cyanobacteria experiments | |
| CN113796873B (en) | Wearable dynamic electrocardiosignal classification method and system | |
| Kulkarni et al. | Depression prediction system using different methods | |
| Ma et al. | The Quality Detection and Recognition for Food Seasoning Based on an Artificial Olfactory System | |
| Çatalbaş | An Investigation into the Relationship between Curse of Dimensionality and Dunning-Kruger Effect | |
| Guoliang et al. | Based on Clustering Algorithm Expert System of Self-learning Rule Base | |
| Butner et al. | Ghost hunting in the nonlinear dynamic machine | |
| Manetta et al. | Verbal representation of fragrances: Dependence on specific task |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200214 |
|
| RJ01 | Rejection of invention patent application after publication |