High-performance gas classification artificial intelligence system and gas classification method using same

Abstract

The present invention provides a high-performance gas classification artificial intelligence system and a gas classification method using same. The high-performance gas classification artificial intelligence system according to the present invention comprises at least one gas sensor, and a gas classification server based on an artificial neural network, wherein the gas classification server includes: a reception unit for receiving gas sensing data from the gas sensor; a pre-processing unit for pre-processing the gas sensing data; and a target gas classification unit for classifying the type and concentration of target gas by inputting the pre-processed gas sensing data into a gas classification model based on the artificial neural network, the gas sensing data is a current signal over time, and the pre-processing unit includes a data processing unit for deriving first differential data and second differential data with respect to the time of the current signal, and calculating processed data by using the first differential data and the second differential data.

Description

High-performance gas classification artificial intelligence system and gas classification method using the same

[0001] The present invention relates to gas classification, and more particularly, to a high-performance gas classification artificial intelligence system and a gas classification method using the same.

[0002] Recently, cases of exposure to toxic gases such as sulfuric acid and carbon monoxide have been rapidly increasing, not only in various industrial settings such as factories and laboratories, but also at fire scenes and in homes. Exposure to these odorless, flammable, and carcinogenic gases can cause illness or death, making it essential to install real-time detection systems capable of detecting and warning of these gases.

[0003] Gas detection systems rely on gas classification technology to accurately detect specific gases. Semiconductor-based micro gas sensors, which feature high discrimination power, high sensitivity, temperature control, on-chip integration, and the ability to detect a wide range of gases, are utilized for gas classification. Semiconductor-based gas sensors output electrical signals when exposed to specific gases, enabling gas classification by identifying the unique electrical signal characteristics of each gas. However, gas sensors age due to changes in time, humidity, temperature, and overexposure to specific gases. This can lead to the detection of the same gas as a different gas, even in the same environment. Therefore, efforts to apply machine learning to gas sensors to classify gases have been ongoing. However, real-world environments often involve a mixture of multiple gases, creating a need for devices and methods that can accurately classify not only single gases but also mixed gases.

[0004] [Prior Art Literature]

[0005] [Patent Document]

[0006] (Patent Document 1) Patent Publication No. 10-2399124 (May 12, 2022)

[0007] The purpose of the present invention is to provide a high-performance gas classification artificial intelligence system capable of accurately classifying the type and concentration of not only a single gas but also a mixed gas through artificial neural network-based machine learning, and a gas classification method using the same.

[0008] The tasks of the present invention are not limited to the tasks mentioned above, and other tasks not mentioned will be clearly understood by those skilled in the art from the description below.

[0009] A high-performance gas classification artificial intelligence system according to one embodiment of the present invention,

[0010] at least one gas sensor; and

[0011] A gas classification server based on an artificial neural network, comprising:

[0012] A receiving unit that receives gas sensing data from the gas sensor;

[0013] A preprocessing unit for preprocessing the above gas sensing data; and

[0014] A target gas classification unit that inputs the preprocessed gas sensing data into the artificial neural network-based gas classification model to classify the type and concentration of the target gas;

[0015] The above gas sensing data is a current signal over time,

[0016] The above preprocessing unit includes a data processing unit that derives first-order differential data and second-order differential data with respect to time of the current signal, and produces processed data using the first-order differential data and the second-order differential data.

[0017]

[0018] Preferably,

[0019] The above-mentioned processed data is characterized in that it is characteristic curve data calculated with the first-order differential data as the X-axis and the second-order differential data as the Y-axis.

[0020]

[0021] Preferably,

[0022] The above target gas classification unit performs classification of the type and concentration of the target gas using the characteristic curve data as input to the gas classification model.

[0023]

[0024] Preferably,

[0025] The above data processing unit,

[0026] A characteristic value extraction unit that extracts slope and curvature values ​​at multiple points from the above characteristic curve data; and

[0027] It includes a feature variable extraction unit that extracts feature variables related to at least one adsorption and desorption of the target gas from the above slope and curvature values.

[0028]

[0029] Preferably,

[0030] The above target gas classification unit performs classification of the type and concentration of the target gas by using characteristic variables related to the adsorption and desorption as inputs to the gas classification model.

[0031]

[0032] Preferably,

[0033] The above target gas is a single gas or a mixed gas,

[0034] When the target gas is the mixed gas, the concentration is characterized by being a mixing ratio of each gas type for the mixed gas.

[0035]

[0036] A gas classification method using a high-performance gas classification artificial intelligence system according to another embodiment of the present invention,

[0037] A gas sensing data receiving step of receiving gas sensing data from at least one gas sensor by a receiving unit;

[0038] A preprocessing step of preprocessing the gas sensing data by a preprocessing unit; and

[0039] A target gas classification step for classifying the type and concentration of the target gas by inputting the gas sensing data preprocessed by the target gas classification unit into a gas classification model based on an artificial neural network;

[0040] The above gas sensing data is a current signal over time,

[0041] The above preprocessing step includes a data processing step of deriving first-order differential data and second-order differential data with respect to time of the current signal by a data processing unit of the preprocessing unit, and producing processed data using the first-order differential data and the second-order differential data.

[0042]

[0043] Preferably,

[0044] The above-mentioned processed data is characterized in that it is characteristic curve data calculated with the first-order differential data as the X-axis and the second-order differential data as the Y-axis.

[0045]

[0046] Preferably,

[0047] The above target gas classification step is characterized in that it performs classification of the type and concentration of the target gas by using the characteristic curve data as input to the gas classification model.

[0048]

[0049] Preferably,

[0050] The above data processing steps are:

[0051] A characteristic value extraction step of extracting slope and curvature values ​​at multiple points from the characteristic curve data by the characteristic value extraction unit of the data processing unit; and

[0052] It includes a feature variable extraction step of extracting a feature variable related to at least one adsorption and desorption of the target gas from the slope and curvature values ​​by a feature variable extraction unit of the data processing unit.

[0053]

[0054] Preferably,

[0055] The above target gas classification step performs classification of the type and concentration of the target gas by using characteristic variables related to the above adsorption and desorption as inputs to the above gas classification model.

[0056]

[0057] Preferably,

[0058] The above target gas is a single gas or a mixed gas,

[0059] When the target gas is the mixed gas, the concentration is characterized by being a mixing ratio of each gas type for the mixed gas.

[0060]

[0061] Specific details of other embodiments are included in the detailed description and drawings.

[0062] The high-performance gas classification artificial intelligence system according to the present invention and the gas classification method using the same have the advantage of being able to accurately classify the type and concentration of not only a single gas but also a mixed gas.

[0063] The high-performance gas classification artificial intelligence system according to the present invention and the gas classification method using the same can perform more accurate gas classification by utilizing characteristics related to gas adsorption and desorption phenomena.

[0064] The high-performance gas classification artificial intelligence system according to the present invention and the gas classification method using the same can implement a high-performance gas classification machine learning system capable of efficiently classifying the type and concentration of a single gas and a mixed gas by linearizing them.

[0065] However, the effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

[0066]

[0067] FIG. 1 is a schematic diagram of a high-performance gas classification artificial intelligence system according to one embodiment of the present invention.

[0068] FIG. 2 is a diagram illustrating an example of sensing data sensed by a gas sensor according to one embodiment of the present invention.

[0069] FIGS. 3A to 3C are drawings illustrating examples of a processing process of a characteristic curve generated by a data processing unit according to one embodiment of the present invention.

[0070] FIG. 4 is a diagram illustrating an example of a modeled function of sensing data according to one embodiment of the present invention.

[0071] FIG. 5 is a drawing schematically explaining extraction of characteristic values ​​such as slope and curvature values ​​by a modeled characteristic curve according to one embodiment of the present invention.

[0072] Figures 6a to 6c are diagrams showing changes in characteristic curves according to changes in characteristic variable values.

[0073] Figures 7a and 7b are enlarged views illustrating the transient region of sensing data sensed by a gas sensor.

[0074] FIG. 8 is a flowchart illustrating a gas classification method using an artificial intelligence system according to one embodiment of the present invention.

[0075] FIG. 9 is a diagram illustrating an exemplary computing device that may implement devices and / or systems according to various embodiments of the present invention.

[0076] Figure 10 is a diagram showing an example of the results of deriving first-order differential data and second-order differential data from gas sensing data received from various commercial gas sensors.

[0077] The advantages and features of the present invention, and the methods for achieving them, will become clearer with reference to the embodiments described in detail below together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below and may be implemented in various different forms. These embodiments are provided only to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. Like reference numerals designate like elements throughout the specification.

[0078] Embodiments described herein will be described with reference to cross-sectional and / or plan views, which are ideal illustrations of the present invention. In the drawings, the thicknesses of components are exaggerated for the purpose of effectively explaining the technical contents. Accordingly, the components illustrated in the drawings have a schematic nature, and the shapes of the components illustrated in the drawings are intended to illustrate specific forms of the components and are not intended to limit the scope of the invention. Although terms such as first, second, and third are used to describe various components in various embodiments of the present specification, these components should not be limited by such terms. These terms are used only to distinguish one component from another. The embodiments described and illustrated herein also include complementary embodiments thereof.

[0079] The terminology used herein is for the purpose of describing embodiments only and is not intended to limit the present invention. In this specification, the singular also includes the plural unless specifically stated otherwise. As used herein, the terms "comprises" and / or "comprising" do not exclude the presence or addition of one or more other components, steps, operations, and / or elements to the mentioned components, steps, operations, and / or elements.

[0080] Unless otherwise defined, all terms (including technical and scientific terms) used herein may be used in their common sense to those of ordinary skill in the art to which the present invention pertains. Furthermore, terms defined in commonly used dictionaries are not to be interpreted ideally or excessively unless explicitly and specifically defined otherwise.

[0081] Hereinafter, with reference to the drawings, the concept of the present invention and embodiments thereof will be described in detail.

[0082]

[0083] FIG. 1 is a schematic diagram of a high-performance gas classification artificial intelligence system according to one embodiment of the present invention.

[0084] A high-performance gas classification artificial intelligence system (100) according to one embodiment of the present invention includes a gas sensor (110), a customer terminal (140), and a gas classification server (120).

[0085] The gas sensor (110), customer terminal (140), and gas classification server (120) can be connected to each other through a network (130).

[0086] The gas sensor (110) in the present invention may be composed of one or more gas sensors, and may be a sensor with an IoT function capable of transmitting sensing data to a receiving unit (121) of a gas classification server (120) via a network (130).

[0087] In the present invention, the gas classification server (120) may refer to a group of servers having a communication interface and capable of executing an artificial neural network including machine learning and / or deep learning.

[0088] The gas classification server (120) includes a receiving unit (121), a preprocessing unit (122), and a target gas classification unit (124).

[0089] As illustrated in Fig. 1, the preprocessing unit (122) may include a data processing unit (123).

[0090] The receiver (121) receives gas sensing data from the gas sensor (110).

[0091] As an example, gas sensing data is measured by a gas sensor (110).

[0092] Types of gas sensors include catalytic sensors, thermal conductivity sensors, semiconductor sensors, infrared absorption sensors, electrochemical sensors, and photoionization detection sensors.

[0093] Depending on the type of sensor, the output of sensing data may vary.

[0094] In one embodiment, the sensing data may be acquired by a semiconductor sensor, and the sensing data may be a current signal over time.

[0095] The preprocessing unit (122) preprocesses gas sensing data received from the gas sensor (110).

[0096] As mentioned above, the preprocessing unit (122) is configured to include a data processing unit (123).

[0097] In addition, the data processing unit (123) includes a feature value extraction unit and a feature variable extraction unit, although not shown in FIG. 1.

[0098] The feature value extraction unit extracts slope and curvature values ​​at multiple points from the feature curve data.

[0099] The feature variable extraction unit extracts feature variables related to at least one adsorption and desorption of the target gas from the slope and curvature values.

[0100] When the gas sensing data is a current signal, the data processing unit (123) derives first-order differential data and second-order differential data with respect to time of the current signal, and produces processed data using the first-order differential data and second-order differential data.

[0101] The processed data is characteristic curve data calculated using the first-order differential data as the X-axis and the second-order differential data as the Y-axis. The characteristic curve is examined in detail using Figures 3a to 3c.

[0102] The present invention is characterized in that it processes gas sensing data into characteristic curve data expressed as 'first-order differential data' and 'second-order differential data' and connects this to machine learning to perform accurate gas classification.

[0103] In conventional gas classification devices using machine learning, analysis was performed using the gas sensing data itself, but the present invention is different in that it processes the gas sensing data into characteristic curve data expressed as 'first-order differential data' and 'second-order differential data' and connects this to machine learning to perform gas classification.

[0104] The target gas classification unit (124) inputs gas sensing data into an artificial neural network-based gas classification model to classify the type and concentration of the target gas.

[0105] The gas classification model according to an embodiment of the present invention is an artificial neural network-based model, and can be prepared in advance by learning with learning data for classifying the type and concentration of target gas.

[0106] In the present invention, the learning data input to the gas classification model of the artificial neural network is the output of the data processing unit (123).

[0107] More specifically, the learning data may be characteristic curve data expressed as ‘first-order differential data’ and ‘second-order differential data’ as output of the data processing unit (123).

[0108] As another example, the learning data may be feature variables related to adsorption and desorption as output of the feature variable extraction unit of the data processing unit (123).

[0109] In an embodiment where the characteristic curve data is learning data, the target gas classification unit (124) performs classification of the type and concentration of the target gas using the characteristic curve data as input to the gas classification model.

[0110] In an embodiment where feature variables related to adsorption and desorption of a target gas are learning data, the target gas classification unit (124) performs classification of the type and concentration of the target gas by using feature variables related to adsorption and desorption of the target gas as input to a gas classification model.

[0111] The target gas classification unit (124) learns gas sensing data as learning data to establish an artificial neural network-based gas classification model, and the learning method may be based on at least one of supervised learning, unsupervised learning, and reinforcement learning.

[0112] Meanwhile, the gas classification model according to the present embodiment may include, but is not limited to, KNN (k-nearest neighborhood), BDT (binary decision tree), MLP (multiple layer perceptron), SVM (support vector machine), RCE (restricted coulomb energy), etc.

[0113] KNN and BDT are primarily applied to linear data. However, as the number of gas sensors increases, data nonlinearity inevitably increases, making them difficult to use in gas applications.

[0114] MLPs can be applied to nonlinear data. However, MLPs not only require long training times, but can also suffer from overfitting, resulting in poor performance. In contrast, SVMs, despite their simple structure, boast high predictive performance and excellent generalization performance. Furthermore, they have been shown to perform best in gas applications.

[0115] MLPs require optimization of network structures such as the number of layers and neurons, while SVMs also require optimization of various learning parameters. Furthermore, both algorithms require complex learning algorithms. In contrast, RCE neural networks have a relatively simple learning method and can dynamically change their network structure based on learning. This allows for learning without any optimization of the network structure. Therefore, compared to MLPs and SVMs, RCE neural networks require relatively short training times, enabling real-time learning. Furthermore, they exhibit high classification performance with low complexity in a variety of fields.

[0116] The RCE neural network is a supervised learning-based algorithm that creates an optimal classifier for each pattern of input data to classify the input feature data.

[0117] The target gas of the present invention may be a single gas or a mixed gas.

[0118] When the target gas is a mixed gas, the target gas classification unit (125) can extract the mixing ratio of gas species for the mixed gas.

[0119] The gas classification server (120) may further include an output unit (not shown). The output unit transmits the type and concentration of the classified target gas to the user terminal (140).

[0120]

[0121] FIG. 2 is a diagram illustrating an example of sensing data sensed by a gas sensor according to one embodiment of the present invention.

[0122] Sensing data may vary depending on the type of sensor, but an example of sensing data that can be acquired by a semiconductor sensor may be a current signal.

[0123] Figure 2 is an example of sensing data, showing a current signal, and it can be confirmed that it repeats periodically.

[0124]

[0125] FIGS. 3A to 3C are drawings illustrating examples of a processing process of a characteristic curve generated by a data processing unit according to one embodiment of the present invention.

[0126] Figures 3a to 3c illustrate characteristic curves representing dynamic responses processed by the data processing unit (124).

[0127] FIG. 3a is a graph of second-order differential data for the current signal sensed by the gas sensor (110) in FIG. 2.

[0128] FIG. 3b is a graph of first-order differential data for the current signal sensed by the gas sensor (110) in FIG. 2.

[0129] The characteristic curve of the present invention is generated using the discrete values ​​of the first-order differential data and the second-order differential data of the sensing data.

[0130] Figure 3c shows the values ​​of the first-order differential data on the X-axis and the second-order differential data on the Y-axis, and shows a characteristic curve in which two circles meet at the origin.

[0131] Based on the distortion direction, degree, and slope of the tangent line at the origin of the two circles of the characteristic curve in Fig. 3c, various features based on the adsorption and desorption principles of each gas are extracted to analyze the dynamic response of the sensing data.

[0132] Although FIG. 3c illustrates an example of a characteristic curve, as an embodiment of the present invention, the data processing unit (124) may be configured to extract various characteristic variables by drawing a surface using first-order differential data and second-order differential data.

[0133]

[0134] FIG. 4 is a diagram illustrating an example of a modeled function of sensing data according to one embodiment of the present invention.

[0135] Comparing the modeled function of Fig. 4 with the sensing data of Fig. 2, it can be seen that the modeled function of Fig. 4 is similar to a graph of one period of the sensing data of Fig. 2.

[0136]

[0137] FIG. 5 is a drawing schematically explaining extraction of characteristic values ​​such as slope and curvature values ​​by a modeled characteristic curve according to one embodiment of the present invention.

[0138] Figure 5 shows the values ​​of the first-order differential data on the X-axis and the second-order differential data on the Y-axis, and depicts a characteristic curve in which two circles meet at the origin.

[0139] The characteristic value extraction unit of the data processing unit (123) extracts slope and curvature values ​​at multiple points from the characteristic curve data.

[0140] Referring to Fig. 5, the slope can be calculated from two straight lines (510, 520) from the origin.

[0141] The curvature value can be calculated at the inflection points (530 and 540) in Fig. 5.

[0142] The feature variable extraction unit of the data processing unit (123) extracts feature variables related to at least one adsorption and desorption of the target gas from the slope and curvature values.

[0143]

[0144] Figures 6a to 6c are drawings showing changes in characteristic curves according to changes in characteristic parameter values.

[0145] Referring to FIGS. 6a to 6c, it can be seen that the shape of the characteristic curves depicted by the first and second differential data of the sensing data is maintained depending on the value of the characteristic variable, but the degree to which the two circles are distorted, i.e. the slope from the origin, etc., changes.

[0146] That is, it can be seen that the values ​​of the characteristic variables related to at least one adsorption and desorption of the target gas derived from the slope and curvature values ​​are related to the slope and curvature values ​​at multiple points of the characteristic curve data.

[0147]

[0148] Figures 7a and 7b are enlarged views illustrating the transient region of sensing data sensed by a gas sensor.

[0149] The present invention is characterized in that it analyzes dynamic reactions regarding adsorption and desorption between gases.

[0150] Dynamic response is the transient response to the injection of a single gas and / or a mixture of gases.

[0151] Figures 7a and 7b show examples of transient responses under different temperature and humidity conditions.

[0152]

[0153] FIG. 8 is a flowchart illustrating a gas classification method using an artificial intelligence system according to one embodiment of the present invention.

[0154] A gas classification method using an artificial intelligence system according to one embodiment of the present invention includes a gas sensing data receiving step of receiving gas sensing data from a gas sensor by a receiving unit (121) in S810.

[0155] S820 includes a preprocessing step of preprocessing gas sensing data by a preprocessing unit (122). In one embodiment, the gas sensing data may be a current signal over time.

[0156] The preprocessing step of S820 includes a data processing step of deriving first-order differential data and second-order differential data with respect to time of a current signal by a data processing unit (123) of the preprocessing unit, and producing processed data using the first-order differential data and second-order differential data.

[0157] In one embodiment, the processed data is characteristic curve data produced using the first-order differential data as the X-axis and the second-order differential data as the Y-axis.

[0158] Fig. 10 is a diagram illustrating an example of the results of deriving first-order differential data and second-order differential data from gas sensing data received from various commercial gas sensors. As illustrated in Fig. 10, when the preprocessing step according to the present embodiment is applied to actual gas sensing data composed of various combinations of commercial gas sensors (MQ6, MQ7, MQ8, MQ9) and various gases, it can be seen that various dynamic response patterns appear. Therefore, through this, learning data for an artificial neural network-based gas classification model described below can be provided.

[0159] In S830, a target gas classification step is included in which the gas sensing data preprocessed by the target gas classification unit (124) is input into a gas classification model based on an artificial neural network to classify the type and concentration of the target gas.

[0160] The target gas of the present invention is a single gas or a mixed gas, and when the target gas is a mixed gas, the target gas classification unit is characterized in that it additionally extracts the mixing ratio of gas species for the mixed gas.

[0161] The target gas classification step is characterized by performing classification of the type and concentration of the target gas using characteristic curve data as input to the gas classification model.

[0162] The data processing step includes a feature value extraction step of extracting slope and curvature values ​​at multiple points from the feature curve data by a feature value extraction step of the data processing unit, and a feature variable extraction step of extracting feature variables related to at least one adsorption and desorption of the target gas from the slope and curvature values ​​by a feature variable extraction step of the data processing unit.

[0163] The target gas classification step performs classification of the type and concentration of the target gas by using characteristic variables related to adsorption and desorption as inputs to the gas classification model.

[0164]

[0165] FIG. 9 is a diagram illustrating an exemplary computing device (900) that may implement devices and / or systems according to various embodiments of the present invention.

[0166] An exemplary computing device (900) capable of implementing devices according to some embodiments of the present disclosure will now be described in more detail with reference to FIG. 9.

[0167] A computing device (900) may include one or more processors (910), a bus (950), a communication interface (970), a memory (930) for loading a computer program (991) executed by the processor (910), and a storage (990) for storing the computer program (991). However, only components related to the embodiment of the present disclosure are illustrated in FIG. 9.

[0168] Accordingly, a person skilled in the art will appreciate that other general components may be included in addition to the components illustrated in FIG. 9.

[0169] The processor (910) controls the overall operation of each component of the computing device (900). The processor (910) may include a central processing unit (CPU), a microprocessor unit (MPU), a microcontroller unit (MCU), a graphics processing unit (GPU), or any other form of processor (910) well known in the art of the present disclosure. In addition, the processor (910) may perform operations for at least one application or program for executing a method according to embodiments of the present disclosure. The computing device (900) may include one or more processors (910). The computing device (900) may refer to artificial intelligence (AI).

[0170] The memory (930) stores various data, commands, and / or information. The memory (930) can load one or more programs (991) from the storage (990) to execute methods according to embodiments of the present disclosure. The memory (930) may be implemented as a volatile memory such as RAM, but the technical scope of the present disclosure is not limited thereto.

[0171] The bus (950) provides communication between components of the computing device (900). The bus (950) may be implemented as various types of buses, such as an address bus, a data bus, and a control bus.

[0172] The communication interface (970) supports wired and wireless Internet communication of the computing device (900). Furthermore, the communication interface (970) may support various communication methods other than Internet communication. To this end, the communication interface (970) may be configured to include a communication module well known in the technical field of the present disclosure.

[0173] According to some embodiments, the communication interface (970) may be omitted.

[0174] Storage (990) can non-temporarily store one or more programs (991) and various data.

[0175] Storage (990) may be configured to include non-volatile memory such as Read Only Memory (ROM), Erasable Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), flash memory, a hard disk, a removable disk, or any form of computer-readable recording medium well known in the art to which the present disclosure pertains.

[0176] The computer program (991) may include one or more instructions that, when loaded into the memory (930), cause the processor (910) to perform methods / operations according to various embodiments of the present disclosure. That is, the processor (910) may perform methods / operations according to various embodiments of the present disclosure by executing one or more instructions.

[0177]

[0178] Although the preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above, and various modifications may be made by a person skilled in the art without departing from the gist of the present invention as claimed in the claims. Furthermore, such modifications should not be understood individually from the technical idea or prospect of the present invention.

Claims

1. At least one gas sensor; and A gas classification server based on an artificial neural network, comprising: A receiving unit that receives gas sensing data from the gas sensor; A preprocessing unit for preprocessing the above gas sensing data; and A target gas classification unit that inputs the preprocessed gas sensing data into the artificial neural network-based gas classification model to classify the type and concentration of the target gas; The above gas sensing data is a current signal over time, A high-performance gas classification artificial intelligence system, comprising: a data processing unit for deriving first-order differential data and second-order differential data with respect to time of the current signal, and generating processing data using the first-order differential data and the second-order differential data.

2. In claim 1, A high-performance gas classification artificial intelligence system, characterized in that the above-mentioned processing data is characteristic curve data calculated using the first-order differential data as the X-axis and the second-order differential data as the Y-axis.

3. In claim 2, A high-performance gas classification artificial intelligence system in which the target gas classification unit performs classification of the type and concentration of the target gas by using the characteristic curve data as input to the gas classification model.

4. In claim 2, The above data processing unit, A characteristic value extraction unit that extracts slope and curvature values ​​at multiple points from the above characteristic curve data; and A high-performance gas classification artificial intelligence system, comprising a feature variable extraction unit that extracts feature variables related to at least one adsorption and desorption of the target gas from the slope and curvature values.

5. In claim 4, A high-performance gas classification artificial intelligence system in which the target gas classification unit performs classification of the type and concentration of the target gas by using characteristic variables related to the adsorption and desorption as inputs to the gas classification model.

6. In claim 1, The above target gas is a single gas or a mixed gas, A high-performance gas classification artificial intelligence system, characterized in that when the target gas is the mixed gas, the concentration is a mixing ratio of each gas type for the mixed gas.

7. A gas sensing data receiving step of receiving gas sensing data from at least one gas sensor by a receiving unit; A preprocessing step of preprocessing the gas sensing data by a preprocessing unit; and A target gas classification step for classifying the type and concentration of the target gas by inputting the gas sensing data preprocessed by the target gas classification unit into a gas classification model based on an artificial neural network; The above gas sensing data is a current signal over time, A gas classification method using a high-performance gas classification artificial intelligence system, wherein the preprocessing step includes a data processing step of deriving first-order differential data and second-order differential data with respect to time of the current signal by a data processing unit of the preprocessing unit, and producing processed data using the first-order differential data and the second-order differential data.

8. In claim 7, A gas classification method using a high-performance gas classification artificial intelligence system, characterized in that the above-mentioned processed data is characteristic curve data calculated using the first-order differential data as the X-axis and the second-order differential data as the Y-axis.

9. In claim 8, A gas classification method using a high-performance gas classification artificial intelligence system, characterized in that the target gas classification step performs classification of the type and concentration of the target gas by using the characteristic curve data as input to the gas classification model.

10. In claim 8, The above data processing steps are: A characteristic value extraction step of extracting slope and curvature values ​​at multiple points from the characteristic curve data by the characteristic value extraction unit of the data processing unit; and A gas classification method using a high-performance gas classification artificial intelligence system, comprising a feature variable extraction step of extracting a feature variable related to at least one adsorption and desorption of the target gas from the slope and curvature values ​​by a feature variable extraction unit of the data processing unit.

11. In claim 10, A gas classification method using a high-performance gas classification artificial intelligence system, wherein the target gas classification step performs classification of the type and concentration of the target gas by using characteristic variables related to the adsorption and desorption as inputs to the gas classification model.

12. In claim 7, The above target gas is a single gas or a mixed gas, A gas classification method using a high-performance gas classification artificial intelligence system, characterized in that when the target gas is the mixed gas, the concentration is a mixing ratio of each gas type for the mixed gas.

Images