Smell search method and smell search system

By using multiple sensor elements and computing devices to generate odor information groups and calculating the sum or difference of different odor information groups, the problem of odor recognition after odor mixing is solved, and accurate odor search and mixing are achieved.

CN113272637BActive Publication Date: 2024-11-26MORI CHIKAN CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN201980087978.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2019-02-08
Publication Date
2024-11-26
Estimated Expiration
2039-02-08

AI Technical Summary

Technical Problem

Existing technologies make it difficult to determine the odor combination when a specific odor is mixed with other odors, and it is also difficult to determine how to mix odors to obtain the desired odor.

Method used

Multiple sensor elements are used to detect odor substances in the gas, generate odor information groups, and the sum or difference of different odor information groups is calculated by a computing device to achieve odor search.

Benefits of technology

It can accurately search for odor combinations after mixing odors or mix specific odors to obtain the desired odor, thus improving the accuracy of odor recognition and mixing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113272637B_ABST
    Figure CN113272637B_ABST
Patent Text Reader

Abstract

The present invention searches for odors obtained when different odors are mixed with each other, or odors mixed with a specific odor in order to obtain a desired odor. A method for searching for odors is provided, which searches for odors based on a plurality of odor information generated using a plurality of sensor elements and a computing device, wherein the plurality of sensor elements output detection signals corresponding to the adsorption conditions by exhibiting inherent adsorption reactions to different odor substances, and the computing device digitizes the detection signals from the plurality of sensor elements to generate a plurality of odor information. The odor searching method includes: a detection step of detecting a first gas containing a plurality of odor substances using a plurality of sensor elements; a generation step of generating a first odor information group generated based on the detection result of the first gas; a preparation step of preparing a second odor information group generated based on the detection result of a second gas; and a calculation step of calculating the sum or difference between the second odor information group and the first odor information group based on the sum or difference of the odor information generated using the same sensor element.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to an odor searching method and an odor searching system for searching odors based on a plurality of odor information. Background Art

[0002] There is known a sensor including a quartz oscillator that specifically adsorbs odor substances in the air in order to measure the odor of the air (see Patent Document 1).

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent Application Laid-Open No. 5-187986 Summary of the invention

[0006] Problems to be solved by the invention

[0007] By using such a sensor, the smell of the air is detected, and the detected signal is digitized to obtain the smell information of each air. However, it is difficult to find out what kind of smell will be obtained when a specific smell is mixed with other smells. In addition, it is difficult to find out what kind of smell is suitable to mix with a specific smell in order to obtain a target smell.

[0008] The present invention has been made in view of the above circumstances, and an exemplary object of the present invention is to provide a smell searching method and a smell searching system for searching for a smell obtained when different smells are mixed with each other, or a smell mixed with a specific smell in order to obtain a desired smell.

[0009] Means for solving problems

[0010] In order to solve the above-mentioned problems, the present invention has the following configuration.

[0011] [1] An odor search method for searching for odors based on a plurality of odor information generated using a plurality of sensor elements and a computing device, wherein the plurality of sensor elements output detection signals corresponding to their adsorption conditions by exhibiting inherent adsorption reactions to different odor substances, and the computing device digitizes the detection signals from the plurality of sensor elements respectively, thereby generating the plurality of odor information as numerical values, wherein the odor search method comprises: a detection step of detecting a first gas containing a plurality of odor substances using the plurality of sensor elements; a generation step of generating a first odor information group consisting of a plurality of odor information generated based on the detection results of the first gas; a preparation step of preparing a second odor information group consisting of a plurality of odor information generated based on the detection results of detecting a second gas different from the first gas using the plurality of sensor elements; and a calculation step of calculating the sum or difference between the second odor information group and the first odor information group based on the sum or difference of the odor information generated using the same sensor elements.

[0012] [2] An odor search system, comprising: an odor sensor having a plurality of sensor elements, and using the plurality of sensor elements to detect a first gas containing a plurality of odor substances, wherein the plurality of sensor elements output detection signals corresponding to the adsorption conditions thereof by exhibiting inherent adsorption reactions to different odor substances; a generating unit, which converts the detection signals from the plurality of sensor elements into digitized values, thereby generating a plurality of odor information as digitized values, and generates a first odor information group consisting of a plurality of odor information generated based on the detection results of the first gas; a preparing unit, which prepares a second odor information group consisting of a plurality of odor information generated based on the detection results of detecting a second gas different from the first gas using the plurality of sensor elements; and a calculating unit, which calculates the sum or difference between the second odor information group and the first odor information group based on the sum or difference of the odor information generated using the same sensor elements.

[0013] Further objects and other features of the present invention will become apparent from the following preferred embodiments described with reference to the accompanying drawings.

[0014] Effects of the Invention

[0015] According to the present invention, it is possible to provide a smell searching method and a smell searching system for searching for a smell obtained when different smells are mixed together or a smell mixed with a specific smell in order to obtain a desired smell. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flowchart showing each step of the odor search method according to the first embodiment.

[0017] Figure 2 It is the detection signal database D1 detected in the detection step S1.

[0018] Figure 3 : is a graph showing the detection signal detected in the detection step S1.

[0019] Figure 4 This is a diagram for explaining the outline of the generation step S2 of the first embodiment.

[0020] Figure 5 This is an example of an image that appears based on the index generated in the generation step S2 of the first embodiment.

[0021] Figure 6 This is an example of the odor information database D2.

[0022] Figure 7 1 is a schematic diagram showing an overview of the odor search system 1 .

[0023] Figure 8 1 is a top view schematically showing the odor sensor 10 .

[0024] Fig. 9 It is schematically indicated Figure 8 Cross-sectional view of the A-A' section. DETAILED DESCRIPTION

[0025] [Implementation Method 1]

[0026] Below, refer to Figure 1 Each step of the odor searching method according to the first embodiment will be described in sequence. Figure 1 This is a flowchart showing each step of the odor search method according to the first embodiment.

[0027] In the first embodiment, the term "smell" refers to a smell that can be acquired by humans or living things including humans as olfactory information, and is a concept including a smell in which a molecular monomer or a molecular group composed of different molecules gathers at respective concentrations.

[0028] In the first embodiment, a substance in which a molecular monomer or a molecular group composed of different molecules constituting the above-mentioned odor is gathered with each concentration is referred to as an "odor substance". However, an odor substance may broadly mean a substance that can be adsorbed to a substance adsorption film of an odor sensor described later in a broad sense. That is, "odor" often includes a plurality of odor substances that are the cause, and there may also be substances that are not recognized as odor substances or unknown odor substances, and therefore, generally, substances that are not considered to be the cause of the odor may also be included.

[0029] The odor search method according to the first embodiment implements the steps of detection step S1, generation step S2, preparation step S3, and calculation step S4 based on the plurality of odor information generated by using the plurality of sensor elements and the computing device. The odor search method according to the first embodiment may further include adjustment step S5, extraction step S6, similarity calculation step S7, and selection step. Each step is described below.

[0030] <Detection Step S1>

[0031] In the detection step S1, a plurality of sensor elements 11 (see Figure 7 ) detects the first gas ( Figure 1 The first gas is the same as other gases, and is a gas containing multiple odor substances. In the detection step S1, in addition to detecting the first gas, other gases such as the second gas may also be detected ( Figure 1 The first gas and the other gas are not detected simultaneously but are detected separately.

[0032] The sensor element 11 is an odor sensor 10 (see Figure 7 ) components, and outputs a detection signal corresponding to the adsorption status by showing a unique adsorption reaction for each different odor substance. That is, the sensor element 11 shows a unique adsorption reaction for each of the multiple odor substances contained in the first gas. By using multiple sensor elements that show a unique adsorption reaction for the first gas to detect the first gas, unique detection signals can be output from the multiple sensor elements 11. Of course, when a gas containing different odor substances is detected, a detection signal different from that when the first gas is detected is output from the multiple sensor elements 11. The specific structure of the sensor element 11 will be described later.

[0033] Figure 2 The detection signal database D1 is detected in the detection step S1. In the detection signal database D1, each sensor element 11 and the detection signal detected in each sensor element 11 are shown in a mutually associated state. Figure 2 In the detection signal database D1 shown in FIG. 1 , the detection signals of the 35 sensor elements 11, 11-01 to 11-35, are stored in a correlated state. Figure 2 In the figure, for the sake of convenience, the description of sensor elements 11-08 to 11-34 is omitted.

[0034] Specifically, the detection signal is the raw data detected by each sensor element 11. When the odor sensor 10 is, for example, a quartz crystal oscillator sensor (QCM), the raw data generated by the sensor element 11 can be set as the time-dependent change in the resonant frequency of the quartz crystal oscillator. That is, the detection signal of the sensor element 11 can be set as the resonant frequency at a plurality of time points at different elapsed times from the start of the operation of the odor sensor 10. For example, Figure 2 As shown, for the detection signal detected in the sensor element 11-01, when the resonant frequency after 0 seconds from the start of the operation of the odor sensor 10 is used as a reference, the resonant frequency detected after 14 seconds is "9.3", and the resonant frequency detected after 16 seconds is "-11.0". In addition, for the detection signal detected in the sensor element 11-02, when the resonant frequency after 0 seconds from the start of the operation of the odor sensor 10 is used as a reference, the resonant frequency detected after 14 seconds from the start of the operation of the odor sensor 10 is "10.7", and the resonant frequency detected after 16 seconds is "-11.7". In addition, at the time of detection, the time interval for recording the detection signal is not particularly limited, and can be set to 1 second intervals, for example.

[0035] The sensor element 11 is preferably detected multiple times, and the average value of the raw data of the multiple detections is used as the detection signal. The number of detections is not particularly limited, and can be set to 3 times, for example. As the average value, an average value based on arithmetic mean (average value) can be used.

[0036] <Generation Step S2>

[0037] In the generating step S2, a first odor information group ( Figure 1 S102). Based on the detection signal of the first gas, the odor information is generated by the computing device. That is, the detection signal output from each sensor element by detecting the first gas is digitized by the computing device, and a plurality of odor information corresponding to each sensor element is generated. Based on the detection signal output by a given sensor element, the odor information corresponding to the given sensor element is generated by the computing device. The specific structure of the computing device will be described later.

[0038] The so-called odor information is a numerical value obtained by the computing device digitizing the detection signal output by the sensor element. That is, the so-called odor information is information obtained by digitizing the odors detected by each sensor element. When n sensor elements are used to detect the first gas, n odor information is obtained. The n odor information obtained by the detection of the first gas are summarized as a first odor information group. The digitization performed by the computing device is preferably a digitization that makes the numerical value 0 (zero) or a positive value, and the odor information is preferably 0 (zero) or a positive value. By setting the odor information to 0 (zero) or a positive value instead of a negative value, the odor can be evaluated more accurately. In addition, by setting the odor information to 0 (zero) or a positive value, it is advantageous when calculating or comparing the sum or difference of odor information.

[0039] The generation step S2 can be executed, for example, through the respective sub-steps of a difference value calculation sub-step S2-1, a logarithmic operation sub-step S2-2, a value classification sub-step S2-3, and an index generation sub-step S2-4.

[0040] <Difference value calculation sub-step S2-1>

[0041] In the differential value calculation sub-step S2-1, for each detection signal detected in the detection step S1, the difference (differential value) between the maximum value and the initial minimum value after the maximum value (hereinafter also referred to as "the minimum value immediately following the maximum value") is calculated. And, in the case where there are multiple differential values ​​(maximum value and the minimum value immediately following it), the value with the largest differential value is used as the differential value of the measurement result. In this way, for each measurement result, a differential value associated with each of the multiple sensor elements 11 is obtained. In addition, by dividing the differential value by the time from the maximum value to the minimum value immediately following the maximum value, the slope of each detection signal can be obtained.

[0042] Figure 3 is a graph showing the detection signal detected in the detection step S1. Figure 3 In the figure, the vertical axis is the displacement amount [Hz] of the resonant frequency detected after a given time based on the resonant frequency 0 seconds after the start of the operation of the odor sensor 10, and the horizontal axis is the elapsed time [seconds] after the start of the operation of the odor sensor 10. Figure 3 , the detection signals about the sensor elements 11 - 01 , 11 - 02 , and 11 - 03 among the detection signals shown in the detection signal database D1 are shown. Figure 3 In FIG. 1 , the detection signal of the sensor element 11-01 is indicated by a solid line, the detection signal of the sensor element 11-02 is indicated by a dotted line, and the detection signal of the sensor element 11-03 is indicated by a dashed line. It is needless to say that the graphs can be generated similarly for the other sensor elements 11-04 to 11-35. Figure 3In the detection signal of sensor element 11-01, the difference value is "22.6 Hz". That is, in the detection signal of sensor element 11-01, it is the difference value between the maximum value "9.3 Hz" at 14 seconds after the start of the operation of the odor sensor 10 and the minimum value "-13.3 Hz" at 17 seconds after the start of the operation of the odor sensor 10.

[0043] When calculating the difference value, the range of the elapsed time after the start of the operation of the odor sensor 10 may be limited. For example, when the measurement of the odor of the sample starts 15 seconds after the start of the operation of the odor sensor 10 and ends 20 seconds after the start of the operation of the odor sensor 10, the range of the elapsed time for calculating the difference value may be set to the period from 14 seconds to 25 seconds after the start of the operation of the odor sensor 10. In addition, the range of the elapsed time can be set arbitrarily.

[0044] <Logarithmic calculation sub-step S2-2>

[0045] In the logarithmic operation sub-step S2-2, the differential values ​​calculated in the differential value calculation sub-step S2-1 are respectively subjected to logarithmic operation to obtain logarithmic values ​​associated with the plurality of sensor elements 11. In the logarithmic operation, the base is not particularly limited, and can be set to 2, for example. In addition, the differential value is the difference between the maximum value and the minimum value, and is a positive value (real number). Similar to the differential value, the logarithmic value of the slope can also be obtained.

[0046] <Value classification sub-step S2-3>

[0047] In the value classification sub-step S2-3, each logarithmic value obtained in the logarithmic operation sub-step S2-2 is classified into a plurality of regions according to the magnitude of the value. The number of regions to be classified is not particularly limited, and can be, for example, 3 to 5 regions. The following describes the case where the classification is into 3 regions.

[0048] In the value classification sub-step S2-3, first determine the maximum logarithmic value and the minimum logarithmic value among the multiple logarithmic values ​​of each sample obtained in the logarithmic operation sub-step S2-2. Next, calculate the quotient when the difference between the maximum logarithmic value and the minimum logarithmic value is divided by 3. Using the quotient obtained in this way, the numerical range between the maximum logarithmic value and the minimum logarithmic value can be divided into three equal areas. That is, it can be divided into three equal areas from the minimum logarithmic value to the value obtained by adding the quotient to the minimum logarithmic value, the value obtained by adding the quotient to the minimum logarithm plus the quotient to the value obtained by adding 2 times the quotient to the minimum logarithm, and the value obtained by adding 2 times the quotient to the maximum logarithmic value.

[0049] Next, the logarithmic values ​​associated with each sensor element 11 are classified into any one of the three regions. For each logarithmic value, a flag for identifying the classified region can also be set. For example, for the three regions obtained by dividing the value into three equal parts, flags can be set in order from small to large values, such as (1), (2), and (3). In this way, the detection signal associated with each sensor element 11 can be classified into three stages according to the size of its value.

[0050] use Figure 4 The above-mentioned generation step S2 will be described in more detail. Figure 4 This is a diagram for explaining the outline of the generation step S2 of the first embodiment. Figure 4 In the table (A), the differential values ​​calculated in the differential value calculation substep S2-1 for a certain sample are shown. The differential values ​​of sensor elements 11-01 to 11-35 are shown respectively. For example, in the table (A), the differential value obtained in the sensor element 11-01 is "38.7", and the differential value obtained in the sensor element 11-02 is "27.0". In addition, for the sake of convenience, the values ​​of sensor elements 11-11 to 11-34 are omitted (the same applies to the table (B) and table (E) described later).

[0051] Next, in the logarithmic operation sub-step S2-2, the difference value for each sensor element 11 is subjected to logarithmic operation. The logarithmic operation here is represented by the following equation (1). That is, the logarithmic operation is performed on the absolute value of the difference value with a base of 2 to obtain a logarithmic value.

[0052] [Logarithmic value] = log2 | [Difference value] | ... Formula (1)

[0053] Table (B) is a table showing logarithmic values ​​related to each sensor element 11 obtained by the logarithmic operation substep S2-2. For example, in Table (B), the logarithmic value calculated based on the differential value obtained in sensor element 11-01 is "5.3", and the logarithmic value calculated based on the differential value obtained in sensor element 11-02 is "4.8".

[0054] Next, the logarithmic values ​​associated with each sensor element 11 are classified into three regions based on the obtained logarithmic values ​​by the value classification substep S2-3. Specifically, first, the largest logarithmic value (maximum value) and the smallest logarithmic value (minimum value) among the logarithmic values ​​associated with each sensor element 11 in the sample being measured are determined. Then, the quotient of the difference between the maximum value and the minimum value divided by 3 is calculated. These determined maximum values, minimum values, and calculated quotients are shown in Table (C). In Table (C), the determined maximum value is "6.7", the determined minimum value is "3.1", and the calculated quotient is "1.2".

[0055] Based on the determined maximum value, minimum value, and calculated quotient, the logarithmic values ​​associated with each sensor element 11 are classified into three stages. When classifying, the classification is performed based on the classification rules shown in Table (D). Specifically, the classification is performed based on the classification rule that the region with the smallest logarithmic value (region 1) is in the range of 3.1≤[logarithmic value]≤4.3, the region with the second smallest logarithmic value (region 2) is in the range of 4.3<[logarithmic value]≤5.5, and the region with the largest logarithmic value (region 3) is in the range of 5.5<[logarithmic value]≤6.7.

[0056] Next, a flag is assigned to each sensor element 11 based on the classification result. Table (E) shows the result of assigning flags to each sensor element 11. Flag (1) is assigned to the sensor element 11 that obtains a logarithmic value corresponding to region 1, flag (2) is assigned to the sensor element 11 that obtains a logarithmic value corresponding to region 2, and flag (3) is assigned to the sensor element 11 that obtains a logarithmic value corresponding to region 3. For example, in Table (E), flag (2) is assigned to sensor element 11-01, flag (1) is assigned to sensor element 11-03, and flag (3) is assigned to sensor element 11-09.

[0057] <Indicator Generation Sub-Step S2-4>

[0058] In the index generation sub-step S2-4, an index is generated based on the logarithmic value (measurement result) classified in the value classification sub-step S2-3, that is, the logarithmic value (detection result) corresponding to each sensor element 11. The index has a value corresponding to each sensor element 11.

[0059] The so-called index is used to obtain the odor image 5 (refer to Figure 5 ) is data based on image data representing the smell of the sample. The index is not data (pixel data) representing information such as the color or position of each pixel forming the smell image 5, but data representing the position or size, color, shape, etc. of the small image 6. The smell image 5 generated based on the index includes a plurality of small images 6 represented by indicators corresponding to the sensor elements 11 respectively. The smell image 5 can be represented by a predetermined display method as a collection of the above-mentioned plurality of small images 6. Each small image 6 can change according to the size of the value of the corresponding index. Specifically, the size, color, and shape of the small image 6 can be changed according to the size of the value of the corresponding index. That is, in the index generation substep S2-4, the index is generated so that the smell of the sample is represented as the smell image 5 in a predetermined display method. In addition, in the index generation substep S2-4, the index is generated in such a way that each small image 6 changes according to the size of the value of the index.

[0060] Figure 5This is an example of an image that appears based on the index generated in the generation step S2 of the first embodiment. Figure 5 The odor image 5 shown is composed of 35 small images 6, each of which is in the shape of a circle. Each small image 6 corresponds to each sensor element 11-01 to 11-35 and is arranged in order from the upper left. Specifically, Figure 5 In the figure, the two small images 6 in the first row from the top correspond to sensor elements 11-01 and 11-02 from the left, and the five small images 6 in the second row from the top correspond to sensor elements 11-03 to 11-07 from the left. In addition, the small image 6 corresponding to sensor element 11-03 (mark (1)) is represented by a small circle, the small image 6 corresponding to sensor element 11-09 (mark (3)) is represented by a large circle, and the small image 6 corresponding to sensor element 11-01 (mark (2)) is represented by a circle of a size between the small circle and the large circle.

[0061] exist Figure 5 In the figure, the shapes of all the small images 6 are represented by circles, but the shapes of the small images 6 are not limited to circles, and may be squares, rectangles, rhombuses, or other irregular shapes. In addition, the shapes of all the small images 6 do not have to be the same, and the small images 6 may have different shapes. Figure 5 In the figure, the color of each small image 6 is represented by black, but the color of each small image 6 is not limited to black, and can be represented by any color. In addition, the colors of all small images 6 do not have to be consistent, and each small image 6 can also be represented by different colors.

[0062] exist Figure 5 In the example, each small image 6 is represented as having a different size according to the value of the corresponding indicator. Specifically, if the value of the indicator is large, the small image 6 is displayed large, and if the value of the indicator is small, the small image 6 is displayed small. Here, the size of each small image 6 can also be classified into multiple stages according to the stages classified in the value classification sub-step S2-3. That is, in the value classification sub-step S2-3, when the three stages of the mark (1), (2), and (3) are classified, the size of the small image 6 can also be classified into three stages for display.

[0063] In a given display mode, the intervals between the small images 6 are preferably constant (arranged at equal intervals). In addition, in a given display mode, the position of each small image 6 (the center or center of gravity of each small image 6) is preferably constant (does not move according to the value of the index). By making the position or interval of each small image 6 constant in this way, when the size or shape of each small image 6 changes according to the value of the index, by comparing the odor image 5 before and after the change, it is possible to visually easily grasp the changed small image 6. In addition, the intervals between the small images 6 are not limited to being constant (equal intervals), and the odor image 5 may also be an image combining a plurality of small images 6 of different shapes.

[0064] <Preparation Step S3>

[0065] In the preparation step S3, the second gas is detected by the sensor element and converted into a numerical value by the computing device in the same manner as the first gas, and a second odor information group ( Figure 1 S104).

[0066] <Calculation Step S4>

[0067] In the calculation step S4, the sum or difference between the first odor information group and the second odor information group is calculated to generate a third odor information group ( Figure 1 The sum or difference between the first odor information group and the second odor information group is calculated based on the sum or difference between odor information generated using the same sensor element.

[0068] When calculating the sum of odor information, one piece of odor information can be directly added to another piece of odor information. Alternatively, one piece of odor information can be multiplied by an arbitrary coefficient and then added to another piece of odor information, or the two pieces of odor information can be multiplied by arbitrary coefficients and then added. When calculating the difference between odor information, one piece of odor information can be directly subtracted from another piece of odor information. Alternatively, one piece of odor information can be multiplied by an arbitrary coefficient and then subtracted from another piece of odor information, or the two pieces of odor information can be multiplied by arbitrary coefficients and then subtracted from one piece of odor information. There is no particular limitation on the coefficient with which the odor information is multiplied, but the odor coefficient can be determined by taking into account the difference in detection intensity between odors caused by the difference in the amount of adsorption of the odor substance to the material adsorption film, etc.

[0069] In this specification, "using the same sensor element" may mean using a single sensor element for detecting both the first gas and the second gas, or using different sensor elements having the same structure for detecting the first gas and the second gas, respectively. It should be noted that when using sensor elements having the same structure to detect gases under fixed conditions, it is a prerequisite to obtain the same detection signal.

[0070] The plurality of sensor elements for detecting the first gas and the plurality of sensor elements for detecting the second gas may be all the same sensor elements or may be partially different sensor elements. The plurality of sensor elements for detecting the first gas and the plurality of sensor elements for detecting the second gas are preferably the same sensor elements. The more the sensor elements used in the detection are the same, the more the number of numerical values ​​obtained increases, so that the odor obtained as a result can be more accurately expressed.

[0071] <Adjustment Step S5>

[0072] In the calculation step S4, sometimes, part of the plurality of odor information in the third odor information group as a result of the difference between the second odor information group and the first odor information group is a negative value. Since the odor information is preferably a positive value, when part of the odor information is a negative value, the part of the odor information can be adjusted to generate the fourth odor information group ( Figure 1 S106).

[0073] As the adjustment step S5 for adjusting the odor information that has a negative value in the third odor information group, for example, the following adjustment steps S5-A to S5-D can be adopted. The adjustment step S5 is not limited to these adjustment steps S5-A to S5-D, and the odor information that has a negative value may be adjusted by other methods.

[0074] (Adjustment step S5-A)

[0075] The adjustment step S5-A is a step of excluding the specific odor information from the third odor information group when the specific odor information included in the third odor information group is a negative value. The third odor information group after excluding the odor information with a negative value can be used as the fourth odor information group (1). The odor information included in the fourth odor information group (1) does not include odor information with a negative value, but only includes odor information with a value of 0 (zero) or a positive value.

[0076] (Adjustment step S5-B)

[0077] The adjustment step S5-B is a step of replacing the value of the specific odor information contained in the third odor information group with 0 (zero) when the value of the specific odor information contained in the third odor information group is negative. The third odor information group after replacing the value of the odor information with 0 (zero) can be used as the fourth odor information group (2). The odor information contained in the fourth odor information group (1) does not include odor information with a negative value, but only includes odor information with a value of 0 (zero) or a positive value.

[0078] (Adjustment step S5-C)

[0079] The adjustment step S5-C is a step of adding the absolute value of the specific odor information to all the odor information contained in the third odor information group when the specific odor information contained in the third odor information group is a negative value. When there are multiple specific odor information with negative values, it is preferred to add the absolute value of the specific odor information with the largest absolute value of the negative value. The third odor information group obtained by adding the largest value among the absolute values ​​of the specific odor information to all the odor information of the third odor information group with specific odor information with negative values ​​can be used as the fourth odor information group (3). The odor information contained in the fourth odor information group (3) does not include odor information with negative values, but only includes odor information with 0 (zero) or positive values.

[0080] (Adjustment step S5-D)

[0081] The adjustment step S5-D is to repeatedly add the first smell information group or the second smell information group to the third smell information group until the specific smell information becomes 0 (zero) or a positive value when the specific smell information included in the third smell information group is a negative value. That is, each smell information of the first smell information group or the second smell information group is repeatedly added to each smell information of the third smell information group so that the specific smell information with a negative value becomes 0 (zero) or a positive value. At this time, when the third smell information group is a calculation result (difference) when the second smell information group is subtracted from the first smell information group, the first smell information group is added to the third smell information group. When the third smell information group is a calculation result (difference) when the first smell information group is subtracted from the second smell information group, the second smell information group is added to the third smell information group. The third smell information group obtained by repeatedly adding the first smell information group or the second smell information group to the third smell information group until the specific smell information with a negative value becomes 0 (zero) or a positive value can be referred to as the fourth smell information group (4). The odor information contained in the fourth odor information group (4) does not include odor information with a negative value, but only includes odor information with a 0 (zero) or positive value.

[0082] <Extraction Step S6>

[0083] In the extraction step S6, a part or all of the plurality of odor information groups 2 are extracted from the odor information database D2 generated in advance (S107). Figure 6 An example of the odor information database D2 is shown. Figure 6 As shown, in the odor information database D2, a plurality of odor information 3 related to a specific odor is stored in association with each other as an odor information group 2. The odor information database D2 stores a plurality of odor information groups 2. Figure 6In FIG. 1 , the difference values ​​calculated for each sensor element 11 are shown for odor a and odor b. For odor a, the difference value calculated using sensor element 11-01 is "38.7", and the difference value calculated using sensor element 11-02 is "27.0". The difference values ​​(odor information) calculated using each sensor element are combined to form odor information group 2 for odor a. In addition, in FIG. Figure 6 In FIG. 1 , for convenience, only the difference values ​​(odor information) of the sensor elements 11 - 01 to 11 - 07 are shown.

[0084] <Similarity calculation step S7>

[0085] In the similarity calculation step S7, the similarity between the odor information group 2 extracted from the database D2 in the extraction step S6 and the third odor information group or the fourth odor information group is calculated (S108). That is, each odor information 3 included in the odor information group 2 extracted from the odor information database D2 in the extraction step S6 and each odor information 3 included in the third odor information group or the fourth odor information group are compared with each other as the odor information 3 respectively corresponding to each other, and the similarity is calculated. The third odor information group is compared with the case where no adjustment is performed in the adjustment step S5. In addition, the so-called "corresponding odor information 3 each other" means the odor information 3 each other detected using the same sensor element.

[0086] When calculating the similarity, if the odor information set 2 is regarded as a vector whose coordinates are each odor information 3 included in the odor information set 2, the similarity between the vectors can be calculated. When the odor information set 2 includes n odor information 3, the vector becomes an n-dimensional vector.

[0087] Specifically, the similarity can be calculated by calculating various indices such as the cosine similarity of two vectors, the cosine similarity (score value) related to the value obtained by dividing the angle θ between the two vectors by π (θ / π) in the cosine similarity, the distance between vectors, the Pearson correlation coefficient, and the deviation pattern similarity. As the similarity, the closer the score value, the cosine similarity, the Pearson correlation coefficient, and the deviation pattern similarity are to 1, the higher the similarity is, and the smaller the value of the distance between vectors is, the higher the similarity is.

[0088] <Selection Steps>

[0089] In the selection step, a specific odor information group 2 is selected from the plurality of odor information groups 2 extracted from the database D2 in the extraction step S6 based on the similarity calculated in the similarity calculation step S7. For example, a odor information group 2 having a similarity with the third odor information group or the fourth odor information group higher than a given threshold value may be selected from the plurality of odor information groups 2 extracted from the database D2, or one or more odor information groups 2 having the highest similarity may be selected.

[0090] [Implementation Method 2]

[0091] Next, the smell search system 1 according to the second embodiment will be described with reference to the drawings. Figure 7 1 is a schematic diagram showing an overview of an odor search system 1. The odor search system 1 includes an odor sensor 10, a generation unit, a preparation unit, and a calculation unit. The generation unit generates a plurality of odor information 3 as numerical values ​​by digitizing the detection signals from the plurality of sensor elements, and generates a first odor information group consisting of the plurality of odor information 3 generated based on the detection result of the first gas. The preparation unit prepares a second odor information group consisting of the plurality of odor information 3 generated based on the detection result of a second gas different from the first gas detected by the plurality of sensor elements. The calculation unit calculates the sum or difference between the second odor information group and the first odor information group based on the sum or difference of the odor information 3 generated using the same sensor element.

[0092] The odor search system 1 may also include: an extraction unit, which extracts part or all of the multiple odor information groups 2 from an odor information database D2 storing multiple odor information groups 2 consisting of multiple odor information 3; a similarity calculation unit, which calculates the similarity between part or all of the odor information groups 2 among the multiple odor information groups 2 extracted by the extraction unit and the third odor information group; and a selection unit, which selects a specific odor information group 2 from the multiple odor information groups 2 based on the similarity calculated by the similarity calculation unit.

[0093] The operation processing device 51 as an operation device can realize the functions of a generating unit, a preparing unit, a calculating unit, an extracting unit, a similarity calculating unit, a selecting unit, etc. The odor search system 1 can realize the above-mentioned odor search method.

[0094] <Odor sensor 10>

[0095] Figure 8 1 is a top view schematically showing the odor sensor 10 . Fig. 9 It is schematically indicated Figure 8The odor sensor 10 includes a plurality of sensor elements 11. Each of the sensor elements 11 includes a substance adsorption film 13 for adsorbing odor substances and a detector 15 for detecting the adsorption state of the odor substances on the substance adsorption film 13.

[0096] like Figure 8 and Fig. 9 As shown, the sensor element 11 is composed of a detector 15 and a substance adsorption film 13 provided on the surface of the detector 15. The substance adsorption film 13 preferably covers the entire surface of the detector 15. That is, the size of the detector 15 is preferably the same as or smaller than the formation range of the substance adsorption film 13. In addition, a plurality of detectors 15 may be provided within the formation range of one substance adsorption film 13.

[0097] The sensor element 11 is provided on the sensor substrate 17 in multiple numbers, such as Figure 8 As shown, six sensor elements 11 can be arranged in a manner to describe an equilateral triangle. In this case, the substance adsorption films 13 of adjacent sensor elements 11 do not contact each other or are insulated. In addition, the sensor elements 11 do not necessarily need to be arranged on the sensor substrate 17, and can also be arranged randomly or in an arbitrary manner. In addition, the sensor elements 11 corresponding to all the odor information 3 constituting one odor information group 2 do not need to be arranged on one sensor substrate 17, and different sensor elements 11 can be arranged on multiple sensor substrates 17.

[0098] It is preferred that the properties of the substance adsorption films 13 of the plurality of sensor elements 11 disposed on the sensor substrate 17 are different from each other. Specifically, it is preferred that all of the plurality of sensor elements 11 are composed of substance adsorption films 13 of different compositions, and that there are no substance adsorption films 13 of the same properties. Here, the properties of the substance adsorption films 13 may also be referred to as the adsorption characteristics of the odorous substance to the substance adsorption films 13. That is, even if it is the same odorous substance (or its aggregate), it exhibits different adsorption characteristics on substance adsorption films 13 with different properties. Figure 8 and Fig. 9 In the figure, for convenience, all the substance adsorption films 13 are shown in the same manner, but in fact, their properties are different from each other. In addition, the adsorption characteristics of the substance adsorption films 13 of each sensor element 11 do not have to be all different, and sensor elements 11 provided with substance adsorption films 13 having the same adsorption characteristics may be provided therein.

[0099] As the material of the substance adsorption film 13, a thin film formed of a π-electron conjugated polymer can be used. In the thin film, at least one of an inorganic acid, an organic acid, and an ionic liquid can be contained as a dopant. By changing the type or content of the dopant, the properties of the substance adsorption film 13 can be changed.

[0100] The π-electron conjugated polymer is not particularly limited, but is preferably a polymer having a π-electron conjugated polymer as a skeleton, such as polypyrrole and its derivatives, polyaniline and its derivatives, polythiophene and its derivatives, polyacetylene and its derivatives, and polyazulene and its derivatives.

[0101] When the π-electron conjugated polymer is in an oxidized state and the skeleton polymer itself is a cation, conductivity can be exhibited by containing anions as a dopant. In the present invention, a neutral π-electron conjugated polymer containing no dopant may be used as the substance adsorption film 13 .

[0102] Specific examples of the dopant include inorganic ions such as chloride ion, chloride oxide ion, bromide ion, sulfate ion, nitrate ion, borate ion, etc.; organic acid anions such as alkylsulfonic acid, benzenesulfonic acid, carboxylic acid, etc.; and polymer acid anions such as polyacrylic acid, polystyrenesulfonic acid, etc.

[0103] Alternatively, a method of chemically balancing doping by allowing a salt such as common salt or an ionic compound containing both cations and anions such as an ionic liquid to coexist with a neutral π-electron conjugated polymer may be used.

[0104] When the state of adding one dopant unit (ion) to every two repeating units constituting the π-electron conjugated polymer is set to 1, the content of the dopant in the π-electron conjugated polymer can be adjusted to a range of 0.01 to 5, preferably to a range of 0.1 to 2. By setting the content of the dopant to a value greater than the minimum value in the range, it is possible to suppress the disappearance of the characteristics of the substance adsorption film 13. In addition, by setting the content of the dopant to a value less than the maximum value in the range, it is possible to suppress the reduction in the effect of the adsorption characteristics of the π-electron conjugated polymer itself, and it is possible to suppress the situation where it is difficult to prepare a substance adsorption film 13 having the desired adsorption characteristics. In addition, since it is usually a film in which the dopant as a low molecular weight substance is dominant, it is possible to suppress the situation where the durability of the substance adsorption film 13 is greatly reduced. Therefore, by setting the content of the dopant to the above range, the detection sensitivity of the odorous substance can be appropriately maintained.

[0105] In order to change the adsorption characteristics of the substance adsorption film 13 in each of the plurality of sensor elements 11, different types of π-electron conjugated polymers may be used. Alternatively, the same type of π-electron conjugated polymer may be used to exhibit different adsorption characteristics by changing the type or content of the dopant. For example, the hydrophobic / hydrophilic properties of the substance adsorption film 13 may be changed by changing the type of π-electron conjugated polymer or the type and content of the dopant.

[0106] The thickness of the substance adsorption film 13 can be appropriately selected according to the characteristics of the odor substance to be adsorbed. For example, the thickness of the substance adsorption film 13 can be in the range of 10 nm to 10 μm, preferably 50 nm to 800 nm. If the thickness of the substance adsorption film 13 is less than 10 nm, sufficient sensitivity may not be obtained. In addition, if the thickness of the substance adsorption film 13 exceeds 10 μm, the upper limit of the weight that can be detected by the detector 15 may be exceeded.

[0107] The detector 15 has a function as a signal conversion unit (converter) that measures changes in physical, chemical, or electrical characteristics of the substance adsorption film 13 caused by the odor substance adsorbed on the surface of the substance adsorption film 13, and outputs the measurement data as, for example, an electrical signal. That is, the detector 15 detects the adsorption state of the odor substance on the surface of the substance adsorption film 13. Examples of the signal output as the measurement data by the detector 15 include physical information such as an electrical signal, light emission, a change in resistance, and a change in vibration frequency.

[0108] The detector 15 is not particularly limited as long as it is a sensor that measures changes in physical, chemical or electrical properties of the substance adsorption film 13, and various sensors can be used as appropriate. Specifically, the detector 15 includes a quartz crystal oscillator sensor (QCM), a surface acoustic wave sensor, a field effect transistor (FET) sensor, a charge coupled device sensor, a MOS field effect transistor sensor, a metal oxide semiconductor sensor, an organic conductive polymer sensor, an electrochemical sensor, and the like.

[0109] In addition, when a quartz oscillator sensor is used as the detector 15, although not shown in the figure, as an excitation electrode, electrodes may be provided on both sides of the quartz oscillator, or a separate electrode may be provided on one side in order to detect a high Q value. In addition, the excitation electrode may also be provided on the sensor substrate 17 side of the quartz oscillator to sandwich the sensor substrate 17. The excitation electrode may be formed of any conductive material. Specifically, as the material of the excitation electrode, inorganic materials such as gold, silver, platinum, chromium, titanium, aluminum, nickel, nickel alloys, silicon, carbon, carbon nanotubes, and organic materials such as conductive polymers such as polypyrrole and polyaniline may be cited.

[0110] like Figure 8 or Fig. 9 As shown, the shape of the detector 15 can be a flat plate. Figure 8 As shown, the shape of the flat plate surface can be circular, but can also be various shapes such as quadrilateral, square, ellipse, etc. The shape of the detector 15 is not limited to a flat plate shape, and its thickness can be varied, and a concave portion or a convex portion can be formed.

[0111] When the detector 15 uses an oscillator like the quartz oscillator sensor described above, the influence (crosstalk) from other oscillators coexisting on the same sensor substrate 17 can be reduced by changing the resonant frequency of each oscillator in the plurality of sensor elements 11. The resonant frequency can be arbitrarily designed so that each oscillator on the same sensor substrate 17 shows different sensitivities to a certain frequency. The resonant frequency can be changed by adjusting the thickness of the oscillator or the substance adsorption film 13, for example.

[0112] As the sensor substrate 17, a silicon substrate, a substrate made of quartz crystal, a printed wiring substrate, a ceramic substrate, a resin substrate, etc. can be used. In addition, the substrate is a multilayer wiring substrate such as an interposer substrate, and the excitation electrodes and mounting wiring for vibrating the quartz substrate and the electrodes for energizing are arranged at arbitrary positions.

[0113] By adopting the above-mentioned structure, it is possible to obtain an odor sensor 10 having a plurality of sensor elements 11, each of which has a substance adsorption film 13 having different adsorption characteristics for an odor substance. Thus, when the odor sensor 10 is used to measure the odor of a certain odor substance or air containing its composition, the odor substance or its composition is in contact with the substance adsorption film 13 of each sensor element 11 in the same manner, but the odor substance is adsorbed to each substance adsorption film 13 in a different manner. That is, the adsorption amount of the odor substance is different in each substance adsorption film 13. Therefore, the detection result of the detector 15 is different in each sensor element 11. Therefore, for a certain odor substance or its composition, the measurement data of the detector 15 corresponding to the number of sensor elements 11 (substance adsorption films 13) provided in the odor sensor 10 is generated.

[0114] The set of measurement data generated by the odor sensor 10 by measuring a certain odor substance or its composition is usually specific (unique) to a specific odor substance or composition of the odor substance. Therefore, by measuring data by the odor sensor 10, the odor can be identified as a single odor substance or a composition (mixture) of odor substances.

[0115] <Computing Device>

[0116] Each unit of the odor search system 1, such as the generation unit, the preparation unit, and the calculation unit, can be realized by the operation processing device 51 as an operation device. Figure 7As shown, the operation processing device 51 is connected to the sensor elements 11 of the odor sensor 10 in a manner that allows communication, and the operation processing device 51 can read the detection results detected by the sensor elements 11 and perform calculations based on the detection results. The operation processing device 51 can also be connected to the storage device 53. In the storage device 53, the program P1 that implements each unit such as the generation unit, the preparation unit, and the calculation unit can be stored. In addition, the detection signal database D1 and the odor information database D2 can be stored in the storage device 53.

[0117] Example

[0118] (Example 1)

[0119] It is known that when the smell of lemon, the smell of lime, and the smell of cinnamon are combined, the smell of cola is obtained. Based on this knowledge, verification was performed using the smell search system 1 of the second embodiment.

[0120] (Sensory test)

[0121] The sensory test was conducted using the odor samples of lemon, lime, and cinnamon from Aromstar. About 10 μl of each odor sample was dripped onto a flavor test paper, and each flavor test paper was sealed in a sample bottle. After leaving it for 10 minutes, the sample bottle was opened and the tester conducted a sensory test. The result of the sensory test was that 7 out of 7 people evaluated it as a flavor close to cola.

[0122] (Example 1-1)

[0123] A lemon odor sample and a cinnamon odor sample manufactured by Aromstar are mixed in advance, and a first odor information group including a difference value of a detection signal is generated for the mixed odor sample. A second odor information group including a difference value of a detection signal is generated for the lime odor sample manufactured by Aromstar. The sum of the first odor information group and the second odor information group is calculated to generate a third odor information group.

[0124] The similarity between the plurality of odor information groups extracted from the odor information database D2 and the third odor information group is calculated, and the plurality of odor information groups extracted are sorted in descending order of similarity, and the first three odor information groups are selected. In addition, the similarity is calculated using cosine similarity, and the sorting is performed based on this. Various conditions of Example 1-1 are shown in Table 1. In addition, the selection results of Example 1-1 are shown in Table 2. In addition, in the odor information database D2, odor information groups generated using the odor sensor 10 for the odors of various samples are stored in association with the labels of each sample. The details of each sample used in the embodiment are shown in Table 3.

[0125] The measurement method of each sample is carried out according to the change of its state. When the state of the sample is liquid, 10 mL of the sample is injected into a small glass bottle with a capacity of 20 mL, and the measurement is performed by bringing the opening of the small glass bottle close to the odor sensor 10. However, when the sample is a fragrance oil, the measurement is performed by bringing the opening of the fragrance oil close to the odor sensor 10. In addition, when the state of the sample is liquid, a few drops of the sample are added to the test paper (perfume paper) and then placed in the sample bottle, and the measurement is performed by bringing the opening of the sample bottle close to the odor sensor 10. When the sample is solid, the sample is placed in a small glass bottle with a capacity of 20 mL, and the measurement is performed by bringing the opening of the small glass bottle close to the odor sensor 10. When the sample cannot be placed in a small glass bottle with a capacity of 20 mL, it is appropriately broken and crushed to the extent that it can be placed in the small glass bottle. When the state of the sample is gas, the cloth exposed to the sample is placed in the sample bottle, and the measurement is performed by bringing the opening of the sample bottle close to the odor sensor 10. When the sample is the exhalation of a smoker, the cloth is exposed to the sample by blowing the exhalation of the smoker during smoking onto the cloth.

[0126] When measuring the odor, an odor measuring device is used which has an odor sensor 10 inside and an inlet for introducing external air to the odor sensor 10. The opening of a small glass bottle or a sample bottle is brought about 1 cm close to the inlet, and the small glass bottle or the sample bottle is kept at this position for about 5 seconds, thereby measuring the odor of the sample.

[0127] [Table 1]

[0128]

[0129] [Table 2]

[0130]

[0131] [Table 3]

[0132]

[0133] As shown in Table 2, in Example 1-1, the smell information group with the highest similarity (first in similarity) to the third smell information group is the smell information group generated for cola beverage A (Pepsi (registered trademark) Cola manufactured by Suntory Foods International Inc.). The smell information group with the second highest similarity is the smell information group generated for black tea A (tea using Earl Grey tea leaves manufactured by Janat), and the smell information group with the third highest similarity is the smell information group generated for aroma oil D (using the aroma oil of "tomato" in the wine aroma set (88 aromas) of Aromstar).

[0134] (Examples 1-2 to 1-10)

[0135] The odor information group was selected in the same manner as in Example 1-1 except that the conditions shown in Table 1 were adopted. In addition, regarding Examples 1-4 to 1-10, when the odor information included in the third odor information group was a negative value, the odor information having a negative value was excluded to generate a fourth odor information group, and the similarity was calculated based on the fourth odor information group. The selection results of Examples 1-2 to 1-10 are shown in Table 2.

[0136] (Example 2)

[0137] It is known that when the smell of isovaleric acid and the smell of vanillin are combined, a chocolate smell is obtained. Based on this knowledge, verification was performed using the smell search system 1 of the second embodiment.

[0138] (Sensory test)

[0139] A sensory test was conducted using a 1000-fold diluted aqueous solution of isovaleric acid and a vanillin odor sample manufactured by Aromstar. A cotton swab to which a small amount of a 1000-fold diluted aqueous solution of isovaleric acid was dripped and a cotton swab to which a odor sample of about 10 mg of vanillin was dripped was sealed in a sample bottle. After leaving it for 10 minutes, the sample bottle was opened and a sensory test was conducted by a tester. As a result of the sensory test, 9 out of 9 people evaluated it as a flavor close to chocolate.

[0140] (Examples 2-1, 3-1 to 3-3)

[0141] The odor information group was selected in the same manner as in Example 1-1 except that the conditions shown in Table 4 were adopted. In addition, for Examples 3-1 to 3-3, when the odor information included in the third odor information group was a negative value, the odor information having a negative value was excluded to generate a fourth odor information group, and the similarity was calculated based on the fourth odor information group. The selection results of Examples 2-1 and 3-1 to 3-3 are shown in Table 5. The details of each sample used in the Examples are shown in Table 3.

[0142] [Table 4]

[0143]

[0144] [Table 5]

[0145]

[0146] Although preferred embodiments of the present invention have been described above, the present invention is not limited thereto, and various modifications and changes can be made within the scope of the gist thereof. For example, the present invention includes the following gist.

[0147] (Subject 1) The first subject is an odor search method, which searches for odors based on multiple odor information generated by using multiple sensor elements and a computing device, wherein the multiple sensor elements output detection signals corresponding to the adsorption conditions by showing inherent adsorption reactions to different odor substances, and the computing device generates the multiple odor information as numerical values ​​by digitizing the detection signals from the multiple sensor elements, wherein the odor search method includes: a detection step of detecting a first gas containing multiple odor substances using the multiple sensor elements; a generation step of generating a first odor information group consisting of multiple odor information generated based on the detection results of the first gas; a preparation step of preparing a second odor information group consisting of multiple odor information generated based on the detection results of detecting a second gas different from the first gas using the multiple sensor elements; and a calculation step of calculating the sum or difference between the second odor information group and the first odor information group based on the sum or difference of the odor information generated using the same sensor elements.

[0148] Thus, it is possible to provide a smell searching method and a smell searching system for searching for a smell obtained when different smells are mixed together or a smell mixed with a specific smell in order to obtain a desired smell.

[0149] (Subject 2) The odor search method may also include an adjustment step, in which, when specific odor information among the multiple odor information possessed by the third odor information group as a result of the difference between the second odor information group and the first odor information group in the calculation step is a negative value, the specific odor information is excluded from the third odor information group.

[0150] (Subject 3) The odor search method may also include an adjustment step, in which, when specific odor information among the multiple odor information possessed by the third odor information group which is the result of the difference between the second odor information group and the first odor information group in the calculation step is a negative value, the value of the specific odor information is replaced with 0.

[0151] (Subject 4) The odor search method may also include an adjustment step, in which, when specific odor information among the multiple odor information possessed by the third odor information group as a result of the difference between the second odor information group and the first odor information group in the calculation step is a negative value, the absolute value of the specific odor information is added to all the odor information possessed by the third odor information group.

[0152] (Subject 5) The odor search method may also include an adjustment step, in which, when specific odor information among a plurality of odor information possessed by a third odor information group as a result of the difference between the second odor information group and the first odor information group in the calculation step is a negative value, the second odor information group is repeatedly added to the third odor information group until the specific odor information becomes 0 or a positive value.

[0153] (Subject 6) The odor search method may further include: an extraction step of extracting a part or all of a plurality of odor information groups from a database storing a plurality of odor information groups consisting of a plurality of odor information;

[0154] a similarity calculating step of calculating the similarity between a part or all of the odor information groups among the plurality of odor information groups extracted by the extracting step and the third odor information group; and

[0155] A selecting step of selecting a specific smell information group from among the plurality of smell information groups based on the similarity calculated in the similarity calculating step.

[0156] (Subject 7) The odor search method may also be configured such that, in the similarity calculation step, with respect to the calculation of the similarity, when the odor information group is set as a vector with each odor information contained in the odor information group as coordinates, any one of the cosine similarity between two odor information groups, the cosine similarity related to the value obtained by dividing the angle θ between the two vectors by π (θ / π) in the cosine similarity, the distance between vectors, the Pearson correlation coefficient, and the deviation pattern similarity is calculated.

[0157] (Subject 8) The odor search method may be configured such that, in the selecting step, a odor information group having the similarity higher than a given threshold value or a odor information group having the highest similarity is selected as the specific odor information group.

[0158] (Subject 9) The second subject is an odor search system, comprising: an odor sensor having a plurality of sensor elements, and using the plurality of sensor elements to detect a first gas containing a plurality of odor substances, wherein the plurality of sensor elements output a detection signal corresponding to the adsorption status by exhibiting an inherent adsorption reaction to different odor substances; a generating unit, which generates a plurality of odor information as numerical values ​​by digitizing the detection signals from the plurality of sensor elements respectively, and generates a first odor information group consisting of a plurality of odor information generated based on the detection result of the first gas; a preparing unit, which prepares a second odor information group consisting of a plurality of odor information generated based on the detection result of detecting a second gas different from the first gas using the plurality of sensor elements; and a calculating unit, which calculates the sum or difference between the second odor information group and the first odor information group based on the sum or difference of the odor information generated using the same sensor elements.

[0159] Explanation of symbols

[0160] 1: Odor search system, 2: Odor information group,

[0161] 3: odor information, 10: odor sensor,

[0162] 11: sensor element, 13: material adsorption film,

[0163] 15: detector, 17: sensor substrate,

[0164] 51: arithmetic processing device, 53: storage device,

[0165] D1: detection signal database, D2: odor information database.

Claims

1. An odor search method, which searches for odors based on a plurality of odor information generated by using a plurality of sensor elements and a computing device, wherein the plurality of sensor elements output detection signals corresponding to their adsorption conditions by exhibiting inherent adsorption reactions to different odor substances, and the computing device converts the detection signals from the plurality of sensor elements into digits, thereby generating the plurality of odor information as numerical values, wherein: The odor searching method comprises: a detection step, using a plurality of the sensor elements to detect a first gas containing a plurality of odor substances; A generating step of generating a first odor information group, wherein the first odor information group is composed of a plurality of odor information generated based on the detection result of the first gas; a preparation step of preparing a second odor information group, the second odor information group being composed of a plurality of odor information generated based on a detection result of a second gas different from the first gas detected by the plurality of sensor elements; and a calculation step of calculating the sum or difference between the second odor information group and the first odor information group based on the sum or difference between the odor information generated using the same sensor element, The odor search method also includes: an adjustment step, in which, when specific odor information among the multiple odor information possessed by the third odor information group as a result of the difference between the second odor information group and the first odor information group in the calculation step is a negative value, the absolute value of the specific odor information is added to all the odor information possessed by the third odor information group.

2. An odor search method, which searches for odors based on a plurality of odor information generated using a plurality of sensor elements and a computing device, wherein the plurality of sensor elements output detection signals corresponding to their adsorption conditions by exhibiting inherent adsorption reactions to different odor substances, and the computing device digitizes the detection signals from the plurality of sensor elements to generate the plurality of odor information as numerical values, wherein: The odor searching method comprises: a detection step, using a plurality of the sensor elements to detect a first gas containing a plurality of odor substances; A generating step of generating a first odor information group, wherein the first odor information group is composed of a plurality of odor information generated based on the detection result of the first gas; a preparation step of preparing a second odor information group, the second odor information group being composed of a plurality of odor information generated based on a detection result of a second gas different from the first gas detected by the plurality of sensor elements; and a calculation step of calculating the sum or difference between the second odor information group and the first odor information group based on the sum or difference between the odor information generated using the same sensor element, The odor search method also includes: an adjustment step, in which, when specific odor information among the multiple odor information possessed by the third odor information group as a result of the difference between the second odor information group and the first odor information group in the calculation step is a negative value, the second odor information group is repeatedly added to the third odor information group until the specific odor information becomes 0 or a positive value.

3. The odor search method according to claim 1 or 2, wherein: The odor search method further comprises: An extraction step of extracting a part or all of a plurality of odor information groups from a database storing a plurality of odor information groups consisting of a plurality of odor information; a similarity calculating step of calculating the similarity between a part or all of the odor information groups among the plurality of odor information groups extracted by the extracting step and the third odor information group; and A selecting step of selecting a specific smell information group from among the plurality of smell information groups based on the similarity calculated in the similarity calculating step.

4. The odor search method according to claim 3, wherein: In the similarity calculation step, the similarity is calculated by setting the odor information group as a vector with each odor information contained in the odor information group as coordinates, and calculating any one of the cosine similarity between two odor information groups, the cosine similarity related to the value θ / π obtained by dividing the angle θ between the two vectors by π, the distance between vectors, the Pearson correlation coefficient, and the deviation pattern similarity.

5. The odor search method according to claim 3, wherein: In the selection step, the odor information group whose similarity is higher than a given threshold or the odor information group with the highest similarity is selected as the specific odor information group.

6. The odor search method according to claim 4, wherein: In the selection step, the odor information group whose similarity is higher than a given threshold or the odor information group with the highest similarity is selected as the specific odor information group.

7. An odor search system for implementing the odor search method of claim 1, comprising: (1) a plurality of sensor elements, each of which exhibits a unique adsorption reaction to a different odor substance, thereby outputting a detection signal corresponding to the adsorption status thereof; (2) a computing device that converts the detection information from the plurality of sensor elements into numerical values ​​to generate a plurality of odor information in the form of numerical values, The plurality of sensor elements are capable of detecting a first gas containing a plurality of odor substances, The computing device functions as the following unit: a generating unit that converts the detection signals from the plurality of sensor elements into numerical values, thereby generating a plurality of odor information as numerical values, and generates a first odor information group consisting of the plurality of odor information generated based on the detection result of the first gas; a preparation unit that prepares a second odor information group consisting of a plurality of odor information generated based on a detection result of a second gas different from the first gas using the plurality of sensor elements; as well as The calculation unit calculates a sum or a difference between the second odor information group and the first odor information group based on a sum or a difference between odor information groups generated using the same sensor element.

8. An odor search system for implementing the odor search method of claim 2, comprising: (1) a plurality of sensor elements, each of which exhibits a unique adsorption reaction to a different odor substance, thereby outputting a detection signal corresponding to the adsorption status thereof; (2) a computing device that converts the detection information from the plurality of sensor elements into numerical values ​​to generate a plurality of odor information in the form of numerical values, The plurality of sensor elements are capable of detecting a first gas containing a plurality of odor substances, The computing device functions as the following unit: a generating unit that converts the detection signals from the plurality of sensor elements into numerical values, thereby generating a plurality of odor information as numerical values, and generates a first odor information group consisting of the plurality of odor information generated based on the detection result of the first gas; a preparation unit that prepares a second odor information group consisting of a plurality of odor information generated based on a detection result of a second gas different from the first gas using the plurality of sensor elements; as well as The calculation unit calculates a sum or a difference between the second odor information group and the first odor information group based on a sum or a difference between odor information groups generated using the same sensor element.