JP7800071B2 - Odor detection device and odor detection system - Google Patents

Description

This disclosure relates to an odor detection system including an odor detection device, and more specifically to detecting deterioration of an odor sensor installed in the odor detection device.

Odor detection devices that determine the type or concentration of an odor from the output values of multiple odor sensors with different characteristics are known. Regarding odor discrimination using multiple odor sensors, for example, International Publication No. 2019/102654 (Patent Document 1) discloses an odor detection device equipped with multiple odor sensors. This odor detection device "includes multiple odor sensors with different odor-reacting characteristics. Based on the output values of the multiple odor sensors, the odor detection device identifies odor components (e.g., nonenal, diacetyl, isovaleric acid) contained in the gas being measured and their concentrations. Based on the identified odor components and their concentrations, the odor type (e.g., aging odor, middle-aged oily odor, sweat odor) is identified, and the identified odor type is output" (see [Abstract]).

Other technologies related to odor discrimination are disclosed, for example, in International Publication No. 2019/102660 (Patent Document 2).

International Publication No. 2019/102654 International Publication No. 2019/102660

The technologies disclosed in Patent Documents 1 and 2 make it impossible to determine which odor sensor has deteriorated when some of the odor sensors among multiple odor sensors have deteriorated. Therefore, there is a need for a technology that can determine which odor sensor has deteriorated when some of the odor sensors among multiple odor sensors have deteriorated.

This disclosure has been made in light of the above-mentioned background, and in one aspect, its purpose is to provide technology for determining which odor sensor has deteriorated when some of multiple odor sensors have deteriorated.

According to one embodiment, an odor detection device is provided. The odor detection device includes three or more odor sensors, a memory unit for storing normal range data obtained by learning the normal range of the output values of each odor sensor for each odor component, and a determination unit for determining degradation of each of the odor sensors. The determination unit compares the output value of each of the odor sensors with the normal range data, and if an odor component is found for which the majority of the actual measured values of each of the odor sensors are within the normal range, determines that the odor sensor that is outside the normal range for the found odor component has degraded.

In an odor detection device according to one aspect, the determination unit determines that the odor components detected by the multiple odor sensors are indistinguishable based on the fact that no odor components are found in the normal range data for which the majority of the actual measured values of each of the multiple odor sensors fall within the normal range.

An odor detection device according to one aspect further includes a correction unit that corrects the actual measurement value of a deteriorated odor sensor. The correction unit compares the actual measurement value of a certain odor sensor with an approximation line of the pre-learned correlation between the concentration of each odor component and the output value of the certain odor sensor to determine whether the deterioration of the certain odor sensor is progressing along the approximation line, and corrects the actual measurement value of the certain odor sensor so that the actual measurement value of the certain odor sensor approaches the approximation line based on whether the deterioration of the certain odor sensor is progressing along the approximation line.

In an odor detection device according to one aspect, the determination unit outputs a determination result indicating that a certain odor sensor has deteriorated based on the difference between the actual measurement value of a certain odor sensor and the output value indicated by the approximation line exceeding a predetermined threshold.

In an odor detection device according to one aspect, the determination unit outputs an estimated time until a certain odor sensor becomes unusable, based on a regression curve obtained from time-dependent data of actual measured values from the odor sensor.

In an odor detection device according to one aspect, the memory unit further stores recorded data of the deterioration time of an odor sensor measured in the past. The determination unit compares the actual measured value of a certain odor sensor with the recorded data and outputs an estimated time until the certain odor sensor becomes unusable.

According to another embodiment, an odor detection system is provided. The odor detection system includes multiple odor detection devices and a communication device. Each of the multiple odor detection devices includes three or more odor sensors, a memory unit for storing normal range data learned from the normal range of output values of each odor sensor for each odor component, and a determination unit for determining deterioration of each of the multiple odor sensors. The determination unit compares the output value of each of the multiple odor sensors with the normal range data, and if an odor component is found for which the majority of the actual measured values of each of the multiple odor sensors are within the normal range, determines that the odor sensor that is outside the normal range for the found odor component has deteriorated. The communication device includes a communication unit for communicating with each of the multiple odor detection devices, and a display unit capable of displaying the results of the deterioration determination obtained from each of the multiple odor detection devices.

In an odor detection system according to one aspect, the determination unit determines that the odor components detected by the multiple odor sensors are indistinguishable based on the fact that no odor components are found in the normal range data for which the majority of the actual measured values of each of the multiple odor sensors fall within the normal range.

In an odor detection system according to one aspect, the odor detection device further includes a correction unit that corrects the actual measurement value of a deteriorated odor sensor. The correction unit compares the actual measurement value of a certain odor sensor with an approximation line of the correlation between the concentration of each odor component, which has been learned in advance, and the output value of the certain odor sensor, to determine whether the deterioration of the certain odor sensor is progressing along the approximation line, and corrects the actual measurement value of the certain odor sensor so that the actual measurement value of the certain odor sensor approaches the approximation line based on whether the deterioration of the certain odor sensor is progressing along the approximation line.

In an odor detection system according to one aspect, the determination unit transmits a determination result indicating that a certain odor sensor has deteriorated to a communication device based on the difference between the actual measurement value of a certain odor sensor and the output value indicated by the approximation line exceeding a predetermined threshold.

In an odor detection system according to a certain aspect, a determination unit transmits to a communication device an estimated time until a certain odor sensor becomes unusable, based on a regression curve obtained from time-dependent data of actual measured values from the odor sensor.

In an odor detection system according to one aspect, the memory unit further stores recorded data of the deterioration time of an odor sensor measured in the past. The determination unit compares the actual measurement value of a certain odor sensor with the recorded data, and transmits to the communication device an estimated time until the certain odor sensor becomes unusable.

In an odor detection system according to one aspect, the communications device further includes a complementing unit that, based on deterioration of the first odor sensor of a first odor detection device included in the plurality of odor detection devices, uses the actual measurement value of the first odor sensor of a second odor detection device included in the plurality of odor detection devices in place of the actual measurement value of the first odor sensor of the first odor detection device to determine a detection result.

According to one embodiment, if some of the multiple odor sensors have deteriorated, it is possible to determine which odor sensor has deteriorated.

These and other objects, features, aspects and advantages of the present disclosure will become apparent from the following detailed description of the disclosure taken in conjunction with the accompanying drawings.

FIG. 1 is a diagram showing an example of a configuration of an odor detection system 10 according to an embodiment. FIG. 10 is a diagram showing an example of a configuration in which a plurality of odor detection devices 100 are used. 10 is a diagram showing an example of the configuration of an odor type determination table 106, and an example of a procedure for identifying odors and determining whether an odor sensor 102 has deteriorated. 1 is a diagram illustrating an example of a process for predicting deterioration of an odor sensor 102 and correcting the odor sensor in an odor detection device 100. FIG. 10 is a flowchart showing an example of a processing procedure for identifying a deteriorated odor sensor 102 in the odor detection system 10.

Hereinafter, embodiments of the technical concepts of the present disclosure will be described with reference to the drawings. In the following description, identical components are designated by the same reference numerals. Their names and functions are also the same. Therefore, detailed descriptions thereof will not be repeated.

FIG. 1 is a diagram showing an example of the configuration of an odor detection system 10 according to the present embodiment. The odor detection system 10 includes an odor detection device 100 and a communication device 120. The odor detection device 100 and the communication device 120 are equipped with short-range wireless communication capabilities such as Bluetooth (registered trademark), and are capable of communicating with each other.

The odor detection device 100 mainly comprises a control unit 101, odor sensors 102A, 102B, 102C, and 102D (collectively referred to as "odor sensors 102"), an ADC (Analog to Digital Converter) 103, a memory unit 104, an operation switch 110, a battery 111, and a communication unit 112. These components can communicate with each other via an internal bus 113. The memory unit 104 also stores an identification unit 105, an odor type determination table 106, a determination unit 107, a correction unit 108, and characteristic data 109.

The control unit 101 can execute programs for realizing the various functions of the odor detection device 100. The control unit 101 is configured, for example, by at least one integrated circuit. The integrated circuit may be configured, for example, by at least one CPU (Central Processing Unit), at least one FPGA (Field Programmable Gate Array), or a combination of these. The control unit 101 may also include RAM such as DRAM (Dynamic Random Access Memory) or SRAM (Static Random Access Memory) for storing programs executed by the CPU. In this case, the control unit 101 can load programs for realizing the various functions of the odor detection device 100 from the storage unit 104 into the RAM and execute the programs.

Odor sensor 102 converts the concentration of the odor component (gas, etc.) to be detected into an electrical signal. In one aspect, odor sensor 102 may be a so-called analog sensor. In this case, odor sensor 102 converts the concentration of the odor component (gas, etc.) to be detected into a voltage or current and outputs it. In another aspect, odor sensor 102 may be a so-called digital sensor. In this case, odor sensor 102 may output the odor component (gas, etc.) to be detected as digital data (e.g., numerical data of approximately 8 to 32 bits). In another aspect, odor sensors 102A, 102B, 102C, and 102D may have different odor response characteristics.

As an example, odor sensor 102 may be realized as a gas sensor for detecting VOCs (Volatile Organic Compounds), a gas sensor for detecting CO, a gas sensor for detecting hydrogen, a gas sensor for detecting hydrocarbons, a gas sensor for detecting alcohol, a gas sensor for detecting tobacco, or a combination of these. Note that each of odor sensors 102A, 102B, 102C, and 102D does not necessarily react to only one type of odor component, but may also react to multiple odor components. Odor components are chemical substances that make up odors.

In this specification, the odor detection device 100 is shown as having four odor sensors 102A, 102B, 102C, and 102D, as an example, but the number of odor sensors 102 included in the odor detection device 100 is not limited to this. In certain aspects, the odor detection device 100 may have any number of odor sensors 102, such as three or more.

ADC 103 converts the analog signals output from odor sensors 102A, 102B, 102C, and 102D into digital signals and outputs them to control unit 101. In some aspects, if odor sensors 102A, 102B, 102C, and 102D are digital sensors, odor detection device 100 may not include ADC 103. In this case, odor sensors 102A, 102B, 102C, and 102D may include an IC (Integrated Circuit) that outputs digital signals.

The storage unit 104 is a non-volatile memory that stores various programs executed by the control unit 101 and various data referenced by the control unit 101. In some aspects, the storage unit 104 may be realized by a hard disk drive (HDD), a solid state drive (SSD), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), flash memory, or the like.

The identification unit 105 is a program for detecting odors or a trained model generated by machine learning. The identification unit 105 is generated, for example, by the following procedure.

First, prepare several different odor components at different concentrations. As an example, prepare odor component X at concentrations 1, 2, and 3, odor component Y at concentrations 1, 2, and 3, and odor component Z at concentrations 1, 2, and 3.

Next, the output values (waveforms) of odor sensors of the same type as odor sensors 102A, 102B, 102C, and 102D (hereinafter referred to as "learning odor sensors") are obtained for each concentration of each odor component. If three types of odor components with three levels of concentration are prepared as described above, the output values X1 to X3 of the learning odor sensors when odor component X is detected at concentrations 1 to 3, the output values Y1 to Y3 of the learning odor sensors when odor component Y is detected at concentrations 1 to 3, and the output values Z1 to Z3 of the learning odor sensors when odor component Z is detected at concentrations 1 to 3 are obtained.

Next, a set of data in which the combination of output values from each training odor sensor is used as input and the odor components and their concentrations are used as output is provided to the learning engine as training data. The learning engine generates a trained model based on the provided training data. This trained model may function as the identification unit 105, or may be used as a library referenced by the identification unit 105.

The above-mentioned "learning odor sensor" is, for example, a sensor having the same characteristics as each of odor sensors 102A, 102B, 102C, and 102D (e.g., odor sensors with the same model number). Furthermore, while the use of a learning odor sensor has been described as an example, machine learning may also be performed using odor sensors 102A, 102B, 102C, and 102D that are actually installed in odor detection device 100.

In some aspects, the above machine learning may use a neural network, in particular a learning vector quantization (LVQ) neural network.

The odor components used for learning can be selected arbitrarily depending on the application of the odor detection system 10. When the application is checking for body odor, the odor components used for learning may be, for example, components that cause malodors, such as nonenal, diacetyl, isovaleric acid, and ammonia. Nonenal is a component that causes body odor associated with aging (body odor that occurs with aging). Diacetyl is a component that causes middle-aged oily odor (a greasy body odor common among middle-aged men). Isovaleric acid is a component that causes sweat odor (body odor caused by sweat).

The odor type discrimination table 106 is a table that associates odor types, concentrations, intensities (odor levels), and the ranges of output values of each of the odor sensors 102A, 102B, 102C, and 102D. The odor type discrimination table 106 is generated during the machine learning process described above. The ranges of output values of each of the odor sensors 102A, 102B, 102C, and 102D are the ranges of expected values that are predicted to be output when a certain odor component is detected by each odor sensor 102 when it is operating normally. In some aspects, the odor type discrimination table 106 may be expressed as a relational database table, or in any other data format such as JSON (JavaScript (registered trademark) Object Notation). The items in the odor type discrimination table 106 will be described later with reference to Figure 3.

In one aspect, odor detection device 100 may input the actual measurement values of each of odor sensors 102A, 102B, 102C, and 102D into identification unit 105, and determine the type, concentration, and intensity of the detected odor based on the trained model. In another aspect, odor detection device 100 may determine the type, concentration, and intensity of the detected odor by comparing the actual measurement values of each of odor sensors 102A, 102B, 102C, and 102D with the range of output values of each of odor sensors 102A, 102B, 102C, and 102D in odor type discrimination table 106.

The odor types may correspond to odor components, or may be human classifications of odors with certain characteristics. As an example, when the odor detection device 100 is used as a body odor checker, the three major body odors of aging, middle-aged oily odor, and sweat odor may be cited as odor types.

By using the odor type determination table 106, the determination unit 107 can also detect partial deterioration of the odor sensors 102A, 102B, 102C, and 102D. Details of detecting partial deterioration of the odor sensors 102A, 102B, 102C, and 102D will be described later with reference to Figure 3.

When some of the odor sensors 102A, 102B, 102C, and 102D have deteriorated, the correction unit 108 corrects the output value of the deteriorated odor sensor. When the odor detection device 100 is used over a long period of time, some of the odor sensors 102A, 102B, 102C, and 102D may deteriorate. For example, when odor sensor 102D deteriorates, the output value of odor sensor 102D may change. This may cause the odor detection device 100 to be unable to correctly distinguish odor components. However, when the degree of deterioration of the deteriorated odor sensor 102D is small (when the amount of change in the output value of odor sensor 102D is small), the effects of deterioration can be suppressed by correcting the output value of odor sensor 102D (e.g., by multiplying by a coefficient). Details of correcting the output value of odor sensor 102 will be described later with reference to Figure 4.

The characteristic data 109 is data indicating the characteristics of each odor sensor 102 referenced by the correction unit 108. By referencing the characteristic data 109, the correction unit 108 can calculate correction values (coefficients) for correcting each odor sensor 102.

The identification unit 105 and the correction unit 108 may be realized as a program executed by the control unit 101. In another aspect, the identification unit 105 and the correction unit 108 may be realized as hardware such as an ASIC (Application Specific Integrated Circuit) having the above-mentioned functions. In this case, the identification unit 105 and the correction unit 108 are provided outside the memory unit 104.

Operation switch 110 may be a power switch for turning the power of odor detection device 100 on/off, a switch for the user to start the odor measurement process on odor detection device 100, or a switch that has both functions. Operation switch 110 may also include multiple switches. In some aspects, odor detection device 100 may be equipped with an LCD monitor and a touch panel.

The battery 111 supplies power to each component of the odor detection device 100. The battery 111 may be a removable dry cell battery or a rechargeable battery. In some aspects, the odor detection device 100 may further include a charging circuit for charging the battery 111 and a power management circuit for managing the voltage and current of the battery 111.

The communication unit 112 transmits and receives data to and from the communication device 120. In one aspect, the communication unit 112 may be implemented as a Bluetooth module and communicate with the communication device 120 in accordance with the BLE (Bluetooth Low Energy) communication standard. In another aspect, the communication unit 112 may be implemented as a wired LAN (Local Area Network) port, a Wi-Fi (registered trademark) (Wireless Fidelity) module, or the like. In yet another aspect, the communication unit 112 may transmit and receive data using a communication protocol such as TCP/IP (Transmission Control Protocol/Internet Protocol) or UDP (User Datagram Protocol).

The communication device 120 may provide various services, such as receiving various information from one or more odor detection devices 100 and displaying it to the user. The communication device 120 may also have a function for determining deterioration of each odor sensor 102 in each odor detection device 100, a function for correcting the output value of each odor sensor 102, and a function for complementing odor sensors 102 between multiple odor detection devices 100. The function for complementing odor sensors 102 between multiple odor detection devices 100 is, for example, a function for, when odor sensor 102A of one odor detection device 100 deteriorates, using the output value of odor sensor 102A of another odor detection device 100 instead. These functions will be described below with reference to FIG. 5. Furthermore, the communication device 120 may store in the storage unit 126 as programs a deterioration determination unit (not shown) for implementing a deterioration determination function for each odor sensor in each odor detection device 100, a correction unit (not shown) for implementing a correction function for the output value of each odor sensor 102, and a complementation unit (not shown) for implementing a complementation function for odor sensors 102 among multiple odor detection devices 100. In this case, the control unit 121 may read each program into RAM and execute it.

Furthermore, the communication device 120 may have the ability to install and execute applications for providing various services to the user. In one aspect, the communication device 120 may be a device owned by the user, such as a personal computer, smartphone, or tablet. In another aspect, the communication device 120 may be a server that collects information from multiple odor detection devices 100. In this case, the communication device 120 may be realized as one or more server devices, or instances (virtual machines) on one or more server devices or in a cloud environment.

The communication device 120 mainly comprises a control unit 121, a display unit 122, an operation unit 123, a first communication unit 124, a second communication unit 125, a memory unit 126, a speaker 127, and a microphone 128.

The control unit 121 may execute programs to realize various functions of the communication device 120. The control unit 121 may be configured, for example, by at least one integrated circuit. The integrated circuit may be configured, for example, by at least one CPU, at least one FPGA, or a combination of these.

The display unit 122 can be realized by a liquid crystal display, an organic EL (Electro-Luminescence) display, or the like.

The operation unit 123 is composed of buttons and/or a touch panel layered on the display unit 122. It outputs operation signals to the control unit 121 according to the position of each button and touch operation.

The first communication unit 124 communicates with external devices connected to the communication network wirelessly or via a wired connection. In some aspects, the first communication unit 124 may be implemented by a wired LAN port, a Wi-Fi module, or the like. In this case, the first communication unit 124 may send and receive data using a communication protocol such as TCP/IP or UDP. The second communication unit 125 functions as an interface for short-range communication such as Bluetooth wireless communication. The communication device 120 can communicate with the odor detection device 100 via either the first communication unit 124 or the second communication unit 125.

The memory unit 126 is a non-volatile memory that stores various programs executed by the control unit 121 and various data referenced by the control unit 121. In certain aspects, the memory unit 126 may be realized by an HDD, SSD, EPROM, EEPROM, flash memory, or the like. The memory unit 126 may also store an application that uses odor data acquired from the odor detection device 100.

The speaker 127 converts electrical signals received from an external device via the first communication unit 124 into audio signals and outputs the audio. The speaker 127 can also output audio when an audio file is played by an application.

The microphone 128 detects sound waves, converts them into electrical signals, and outputs them to the control unit 121 or the first communication unit 124.

Figure 2 shows an example of a configuration when multiple odor detection devices 100 are used. The communication device 120 can communicate with any number of odor detection devices 100, including one or more. The number of odor detection devices 100 required varies depending on the services provided by the communication device 120 or the applications executed by the communication device 120.

As one example, in an application such as a body odor checker, the communication device 120 only needs to communicate with one odor detection device 100. As another example, if the communication device 120 provides a safety confirmation service or disaster notification service used in a home, nursing home, factory, or laboratory, the communication device 120 may communicate with multiple odor detection devices 100 located in a home or nursing home. The communication device 120 may also function as a server and provide services to multiple terminals using information aggregated from the odor detection device 100.

Figure 3 shows an example of the configuration of the odor type determination table 106, and the procedure for identifying odors and determining whether the odor sensor 102 has deteriorated. The procedure for determining whether the odor sensor 102 has deteriorated in the odor detection device 100 will be described with reference to Figure 3 .

In the first step, each of odor sensors 102A, 102B, 102C, and 102D detects or measures measurement target gas 301. Then, each of odor sensors 102A, 102B, 102C, and 102D outputs an output value (hereinafter sometimes referred to as an "actual measurement value") that is the measurement result of measurement target gas 301 to identification unit 105. In one aspect, this processing can be realized by control unit 101 executing a program that outputs an actual output signal to identification unit 105. In another aspect, odor detection device 100 may be equipped with a hardware function that outputs an actual output signal to identification unit 105.

In the second step, the identification unit 105 identifies the type, concentration, and intensity of the measurement target gas 301. As an example, the identification unit 105 of the odor detection device 100 may input the actual measurement values of each of the odor sensors 102A, 102B, 102C, and 102D into the trained model, thereby obtaining the type, concentration, and intensity of the measurement target gas 301 as the output of the trained model.

As another example, the odor detection device 100 or the identification unit 105 may determine the type, concentration, and intensity of the detected odor by comparing the actual measurement values of each of the odor sensors 102A, 102B, 102C, and 102D with the range of output values of each of the odor sensors 102A, 102B, 102C, and 102D in the odor type discrimination table 106. In this case, the odor detection device 100 compares the actual measurement values of each odor sensor 102 with the range of output values of each odor sensor 102 in each record of the odor type discrimination table 106.

The range of output values of each odor sensor 102 is the range of output values that each odor sensor 102 is expected to output when it observes gas of that record type and concentration under normal conditions (hereinafter, this may be simply referred to as the "range of normal values"). In one aspect, the range of normal values may be determined based on, for example, machine learning training data obtained through experiments. In another aspect, the range of normal values may be determined based on the results of having the identification unit 105 identify odor components by concentration (based on the output results of the identification unit 105 when each odor sensor 102 is normal).

The identification unit 105 may select as the measurement result the "type" and "concentration" of the gas in the record 302 where the normal value range of each odor sensor 102 includes the majority of the actual measured values of each odor sensor 102. The identification unit 105 outputs a signal indicating the selected measurement result to the control unit 101. The signal indicating the measurement result may include the type, concentration, and level of the measurement target gas 301, and the actual measured values of each odor sensor 102. In the example of FIG. 3, the identification unit 105 selects "type: X" and "concentration: 1 ppm < x ≦ 5 ppm" as the measurement result.

In some cases, the identification unit 105 may select as the identification result a record having a normal range that includes the largest number of combinations of actual measurement values. For example, suppose the actual measurement values of odor sensor 102A are "a3 < a ≦ a4," the actual measurement values of odor sensor 102B are "b3 < b ≦ b4," the actual measurement values of odor sensor 102C are "c3 < c ≦ c4," and the actual measurement values of odor sensor 102D are "d5 < d ≦ d6." In this case, record 302 includes the actual output signals of odor sensors 102A, 102B, and 102C within the normal range, resulting in the highest degree of match. Therefore, the identification unit 105 selects record 302 ("Type: X," "Concentration: 1 ppm < x ≦ 5 ppm," "Level: 3") as the identification result and outputs a signal indicating the identification result to the control unit 101. Furthermore, if there is no record with a normal range that includes a majority of the actual measurement values or a certain number or more of the actual measurement values (e.g., 3/4 or more of the sensors) the identification unit 105 may output a signal indicating identification failure to the control unit 101. In this way, by using the odor type discrimination table 106, the odor detection device 100 can identify the measurement target gas 301 even if the actual measurement values deviate slightly from the machine learning learning results due to deterioration of some of the odor sensors 102, for example.

In the third step, the determination unit 107 acquires the measurement results (which may include the type, concentration, and level of the measurement target gas 301, as well as the actual measurement values of each odor sensor 102) via the control unit 101 or directly from the identification unit 105. The determination unit 107 then compares the measurement results with the odor type discrimination table 106 to determine whether or not a degraded odor sensor 102 is present, and, if a degraded odor sensor 102 is present, identifies which odor sensor 102 has deteriorated.

More specifically, the determination unit 107 first compares the actual measurement values of each odor sensor 102 with the normal range of each odor sensor 102 to determine whether a record exists in which the majority of the actual measurement values of each odor sensor 102 fall within the normal range. For example, suppose the actual measurement values of odor sensor 102A are "a3 < a ≦ a4," the actual measurement values of odor sensor 102B are "b3 < b ≦ b4," the actual measurement values of odor sensor 102C are "c3 < c ≦ c4," and the actual measurement values of odor sensor 102D are "d5 < d ≦ d6." In this case, record 302 contains the majority of the actual measurement values of odor sensors 102A, 102B, and 102C (three-quarters of the actual measurement values) within the normal range. Therefore, the determination unit 107 determines that there is a record 302 in which the majority of the actual measurement values of each odor sensor 102 are within the normal range, and outputs a signal indicating this determination result to the control unit 101.

Next, the odor sensor 102 that output an actual measurement value that is outside the normal range of record 302 is identified. In this example, the actual measurement value of odor sensor 102D, "d5<d≦d6," is not included in the normal range of odor sensor 102D, "d3<d≦d4," in record 302. In this case, the determination unit 107 determines that odor sensor 102D has deteriorated. This is because the majority of odor sensors 102A, 102B, and 102C are normal because they output actual measurement values that are within the normal range in record 302, and it is highly likely that the minority of odor sensors 102D that output actual measurement values that are outside the normal range in record 302 have deteriorated.

As another example, if there are seven odor sensors 102, and the actual measurement values of five of the odor sensors 102 are within the normal range of a certain record, and the actual measurement values of the remaining two odor sensors 102 are outside the normal range of a certain record, the determination unit 107 may determine that the two odor sensors 102 (minority) that output actual measurement values outside the normal range have deteriorated.

As another example, if the number of odor sensors 102 is even, and the actual measurement values of half of the odor sensors 102 are within the normal range of a certain record, and the actual measurement values of the remaining half of the odor sensors 102 are outside the normal range of a certain record, the determination unit 107 will not be able to determine which odor sensors 102 (the minority odor sensors 102) have deteriorated, and will therefore output a signal indicating an inability to determine or an error to the control unit 101.

In one aspect, the identification unit 105 and the determination unit 107 may be implemented as a single program module or hardware. In another aspect, the second and third steps may be executed simultaneously. Furthermore, in another aspect, the communication device 120 may include the identification unit 105, the odor type determination table 106, and the determination unit 107, and may execute the processing described with reference to FIG. 3. In this case, one or more odor detection devices 100 communicating with the communication device 120 transmit the actual measurement values of each odor sensor 102 to the communication device 120. The communication device 120 may compare the actual measurement values of each odor sensor 102 received from each odor detection device 100 with the normal range of the odor type determination table 106, and identify the measurement results of each odor detection device 100 and any degraded odor sensors 102.

Figure 4 is a diagram showing an example of the deterioration prediction of the odor sensor 102 and the correction process for the odor sensor 102 in the odor detection device 100. The correction process for the odor sensor 102 and the deterioration prediction of the odor sensor 102 in the odor detection device 100 will be described with reference to Figure 4 .

Graph 401 shows characteristic data 109 of odor sensor 102D. Graph 401 is an approximate line calculated from graph 402, which was obtained through an experiment using odor sensor 102D or an odor sensor with the same characteristics (or model number) as odor sensor 102D. That is, graphs 401 and 402 show the correlation between the concentration of each pre-learned odor component and the output value of each odor sensor 102. Graph 401 shows the feature quantity (output value) output by odor sensor 102D for each concentration of a given odor component. In addition to the response value obtained from the sensor, processed data obtained from the sensor can also be used as a feature quantity. For example, the slope of the waveform obtained by plotting the time elapsed since the start of measurement, the peak value, the difference from the initial value, etc. can be used as a feature quantity. In some aspects, characteristic data 109 may exist for each odor sensor 102, the number of which is equal to the number of odor components to be detected. For example, if there are four odor sensors 102, four types of feature amounts, and four types of odor components to be detected, there can be 64 pieces of characteristic data 109 (4*4*4).

In the example of FIG. 4, when the concentration of a certain odor is "5 ppm," odor sensor 102D is expected to output a feature (output value) of "d4." When the concentration of a certain odor is "5 ppm," if the actual measurement value of odor sensor 102D is "d4'," i.e., if the actual measurement value of odor sensor 102D is not "d4," odor sensor 102D may be degraded.

When the amount of deterioration of odor sensor 102D (the amount of change in output value from the expected value) is equal to or less than predetermined threshold 403, correction unit 108 determines that the amount of deterioration of odor sensor 102D is within a correctable range and corrects the output value of odor sensor 102D. In the example of Figure 4, threshold 403 is "d4limit" when the concentration of a certain odor is "5 ppm." When the amount of deterioration of odor sensor 102D (the amount of change in output value from the expected value) is equal to or less than predetermined threshold 403, this means that the amount of deterioration of odor sensor 102D is somewhat in line with graph 401, which shows the characteristics of odor sensor 102D, and is therefore correctable. Conversely, when the amount of deterioration of odor sensor 102D (the amount of change from the expected output value) is equal to or exceeds predetermined threshold 403, this indicates that the amount of deterioration of odor sensor 102D deviates significantly from graph 401, which shows the characteristics of odor sensor 102D, and is uncorrectable (has deteriorated to the point where correction is not possible).

When the amount of deterioration of odor sensor 102D is equal to or less than predetermined threshold 403 or less than threshold 403 (when the degree of deterioration of odor sensor 102 is small), correction unit 108 corrects the output value of odor sensor 102. In the example of Figure 4, when the concentration of a certain odor is "5 ppm" and the actual measured value of odor sensor 102D is "d4'", the correction value (coefficient) can be, for example, "d4/d4' = (δx + α)/d4'".

If the amount of deterioration of odor sensor 102D is equal to or exceeds predetermined threshold 403 (if the degree of deterioration of odor sensor 102 is high), correction unit 108 determines that the output value of odor sensor 102 cannot be corrected. Based on the determination that the output value of odor sensor 102 cannot be corrected, correction unit 108 outputs a signal to control unit 101 indicating that the output value of odor sensor 102 cannot be corrected or that odor sensor 102 has deteriorated.

The correction unit 108 similarly calculates correction values or performs deterioration assessments for the other odor sensors 102A, 102B, and 102C. In some cases, the correction unit 108 may calculate correction values or perform deterioration assessments for each odor sensor 102 using graph 402 obtained through experiments, rather than graph 401.

In one aspect, the memory unit 104 may further store recorded data of the deterioration time measured in the past for each odor sensor 102. The determination unit 107 or correction unit 108 may then compare the actual measurement value of a certain odor sensor 102 with the recorded data of that odor sensor 102, predict the estimated time until the odor sensor 102 becomes unusable (the timing at which the actual measurement value of the odor sensor 102 will be equal to or exceed the threshold value 403), and output the prediction result to the control unit 101.

The correction unit 108 may also periodically acquire actual measurement values (time-dependent data) of each odor sensor 102, calculate a regression curve from changes in the actual measurement values, and predict odor sensor 102 failure (calculate the estimated time until the odor sensor 102 becomes unusable) based on the regression curve. In the example of FIG. 4, the correction unit 108 predicts the time when odor sensor 102D will become unusable (the timing when the actual measurement value of odor sensor 102D will be equal to or exceed threshold value 403) based on regression curve 404 of the deterioration prediction of odor sensor 102D.

As described above, the odor detection device 100 calculates correction values for deteriorated odor sensors 102 based on the characteristic data 109 of each odor sensor 102. This allows the odor detection device 100 to correct the actual measurement values of odor sensors 102 with a low degree of deterioration. The odor detection device 100 can also notify the communication device 120 of information on odor sensors 102 with a high degree of deterioration. Furthermore, the odor detection device 100 calculates a regression curve for predicting deterioration from the time-series actual measurement values of each odor sensor 102. This allows the odor detection device 100 to output the predicted lifespan of the odor sensor 102 to the control unit 101. In certain aspects, the communication device 120 may have the functions described with reference to FIG. 4. In this case, the communication device 120 includes the correction unit 108 and the characteristic data 109 as hardware or software. Furthermore, the communication device 120 can perform correction processing for each odor sensor 102 in each odor detection device 100 and predict deterioration of the odor sensor 102 based on the actual measurement values of each odor sensor 102 obtained from each of one or more odor detection devices 100.

In one aspect, the communication device 120 may receive various data from each odor sensor 102 and control the deterioration determination and correction process for each odor sensor 102 in each odor detection device 100. In this case, each odor detection device 100 transmits to the communication device 120 information such as whether the deterioration of a certain odor sensor 102 has exceeded the threshold 403, the time when the odor sensor 102 will become unusable (the timing when the actual measurement value of the odor sensor 102 will be equal to or exceed the threshold 403), etc. Based on this received information, the communication device 120 may correct the actual measurement value received from each odor detection device 100, or may send a command to each odor detection device 100 to correct the actual measurement value.

Figure 5 is a flowchart showing an example of a processing procedure for identifying a deteriorated odor sensor 102 in the odor detection system 10. In one aspect, the control unit 101 may load a program for performing processing in the odor detection device 100 of Figure 5 from the storage unit 104 into RAM and execute the program. In another aspect, the control unit 121 may load a program for performing processing in the communication device 120 of Figure 5 from the storage unit 126 into RAM and execute the program. Furthermore, in another aspect, some or all of the processing may be realized as a combination of circuit elements configured to perform the processing.

The following process will be described assuming that the odor detection system 10 includes three odor detection devices 100A, 100B, and 100C, and a communication device 120. Note that the following description is an example, and the odor detection system 10 may include any number of odor detection devices 100. Furthermore, it is assumed that odor detection device 100A, odor detection device 100B, and odor detection device 100C are installed in room X to detect or measure the gas concentration in room X (collectively referred to as "odor detection device 100").

In step S505, the odor detection device 100 detects odor components and concentrations.

In step S510, the odor detection device 100 compares the data (normal range) in the odor type determination table corresponding to the detection results (type, concentration, and level of odor component) with the sensor output (actual measured value). The odor detection device 100 may transmit the comparison results (including information on which odor sensor 102 output values are within the normal range or not) to the communication device 120. In some aspects, the communication device 120 may execute the processing of step S510. In this case, the communication device 120 receives the odor component and concentration detection result data from each odor detection device 100 in step S505.

In step S515, the communication device 120 determines whether there is an odor detection device 100 in which all odor sensors 102 are normal. When all odor sensors 102 in a certain odor detection device 100 are normal, it means that the actual measurement values of all odor sensors 102 in that odor detection device 100 are within the normal range of one of the records in the odor type determination table 106. For example, if all of the actual measurement values of odor sensors 102A, 102B, 102C, and 102D of odor detection device 100A are within the normal value range of record 302, then all odor sensors 102 in odor detection device 100A are normal.

If the communication device 120 determines that there is an odor detection device 100 in which all odor sensors 102 are normal (YES in step S515), it transfers control to step S520. Otherwise, the communication device 120 transfers control to step S530. In one aspect, the communication device 120 may receive actual measurement values of the odor sensors 102 from the odor detection device 100 and determine whether each odor sensor 102 of each odor detection device 100 is normal. In another aspect, the communication device 120 may receive data from each odor detection device 100 indicating whether each odor sensor 102 is normal.

In step S520, the communication device 120 detects odor components and concentrations for an odor detection device 100 in which all odor sensors are normal. More specifically, the communication device 120 uses the actual measurement data or judgment results (type, concentration, and level of odor components) received from an odor detection device 100 in which all odor sensors are normal. For example, when detecting the gas concentration in room X, assume that all odor sensors 102 in odor detection device 100A are normal, but some of the odor sensors 102 in odor detection device 100B are degraded. In this case, the communication device 120 uses the actual measurement data or judgment results received from odor detection device 100A.

In step S525, the communication device 120 outputs the results. In one aspect, the communication device 120 may output the results to the display unit 122. In another aspect, the communication device 120 may transmit the results to another device via the first communication unit 124 or the second communication unit 125.

In step S530, the communications device 120 stores the odor sensor 102 whose output value is outside the normal range. For example, suppose the odor detection device 100 or communications device 120 determines that the odor component detected in step S505 is of type "X" and level "3" (corresponding to record 302). Then, suppose the output value of odor sensor 102D in odor detection device 100A is outside the normal range of record 302. In this case, the communications device 120 stores odor sensor 102D in odor detection device 100A as the odor sensor 102 whose output value is outside the normal range.

In step S535, the communication device 120 determines whether it is possible to complement a deteriorated odor sensor 102 among multiple odor detection devices 100. Complementing an odor sensor 102 means, for example, using the odor sensor 102A of odor detection device 100B or the odor sensor 102A of odor detection device 100C in place of the deteriorated odor sensor 102A of odor detection device 100A. This determination can be made based on, for example, the distance between each of the odor detection devices 100A, 100B, and 100C, and the overlap of the inspection areas of each odor detection device 100 (whether both odor detection devices 100A and 100B have the same room as their inspection area, etc.).

If the communication device 120 determines that it is possible to complement a deteriorated odor sensor 102 among multiple odor detection devices 100 (YES in step S535), it transfers control to step S540. If not (NO in step S535), the communication device 120 transfers control to step S545.

In step S540, the communication device 120 redetects the odor components and concentrations using the interpolated output values. For example, the communication device 120 redetects the odor components and concentrations using the output values of the odor sensor 102A of the odor detection device 100B instead of the output values of the deteriorated odor sensor 102A of the odor detection device 100A.

In step S545, the communication device 120 counts the number of odor sensors 102 whose output values are outside the normal range. The communication device 120 performs this count for each odor detection device 100. For example, suppose that odor sensor 102A of odor detection device 100A is degraded. Furthermore, suppose that odor sensors 102A and 102B of odor detection device 100B are degraded. In this case, the count for odor detection device 100A becomes "1" and the count for odor detection device 100B becomes "2".

In step S550, the communication device 120 determines, for each odor detection device 100, whether the proportion of odor sensors whose output values are outside the normal range is half or more of all the odor sensors. For example, odor detection device 100A has four odor sensors 102A, 102B, 102C, and 102D. Suppose that the output values of odor sensors 102A and 102B are outside the normal range. In this case, half of all the odor sensors in odor detection device 100A have output values outside the normal range.

If the communication device 120 determines that, for each odor detection device 100, half or more of all odor sensors have output values outside the normal range (YES in step S550), it transfers control to step S555. If not (NO in step S550), the communication device 120 transfers control to step S560.

In step S555, the communication device 120 outputs a determination error. This determination error is made for each odor detection device 100. For example, if more than half of the odor sensors in both odor detection devices 100A and 100B have output values outside the normal range, the communication device 120 outputs a determination error for odor detection device 100A and a determination error for odor detection device 100B.

In one aspect, the communication device 120 may output the determination error to the display unit 122, or may transmit the determination error to another device via the first communication unit 124. In another aspect, each odor detection device 100 may output the determination error to the communication device 120. In this case, the communication device 120 outputs the received determination error to the display unit 122 or the first communication unit 124.

In step S560, the communication device 120 calculates a correction coefficient for the deteriorated odor sensor for each odor detection device 100. The communication device 120 calculates the correction coefficient using the procedure described with reference to FIG. 4. In some aspects, the odor detection device 100 may calculate the correction coefficient for the odor sensor. In this case, the odor detection device 100 transmits the corrected output value to the communication device 120.

In step S565, the communication device 120 determines whether the deterioration of the odor sensor 102 for each odor detection device 100 has reached the correction limit value. If the communication device 120 determines that the deterioration of the odor sensor 102 has reached the correction limit value (YES in step S565), it transfers control to step S555. If not (NO in step S565), the communication device 120 transfers control to step S570.

In step S570, the odor detection device 100 multiplies the output value to be corrected received from the odor detection device 100 by the correction coefficient calculated in step S560. In some aspects, the odor detection device 100 may transmit the corrected output value to the communication device 120.

In some cases, if the processing of steps S535 and S540 (correction processing between multiple odor detection devices 100) is not required, all of the processing in Figure 5 may be performed by a single odor detection device 100.

In another aspect, the odor detection device 100 may have both or either a function to independently execute the processing of FIG. 5 excluding the processing of steps S535 and S540, and a function to execute the processing of FIG. 5 in cooperation with the communication device 120 and other odor detection devices 100.

Furthermore, in other aspects, the processes in FIG. 5 , excluding the processes of steps S535 and S540, may be executed by either the communication device 120 or the odor detection device 100. When each process is executed on the communication device 120 side, various correction processes and judgment processes may be executed on the actual measurement values obtained from each odor detection device 100. When each process is executed on the odor detection device 100 side, the odor detection device 100 may execute various correction processes and judgment processes on the actual measurement values of its own device, and output the results of each process (corrected output values, deterioration judgment results of the odor sensor 102, etc.) to the communication device 120 together with the actual measurement values of the odor components.

As explained above, the odor detection system 10 or odor detection device 100 according to this embodiment compares the normal range of each odor sensor 102 contained in the odor type discrimination table 106 with the actual measurement value of each odor sensor 102. Then, if the actual measurement values of the majority of odor sensors 102 are within the normal range, the odor detection system 10 or odor detection device 100 can determine that the odor sensors 102 whose actual measurement values are not within the normal range are degraded.

In addition, in certain aspects, when the degree of deterioration of the odor sensor 102 is small, the odor detection system 10 or odor detection device 100 can suppress the effect of deterioration of the odor sensor 102 by multiplying the actual measurement value by a complementary coefficient.

In another aspect, the odor detection system 10 or odor detection device 100 can predict when the odor sensor 102 will fail based on the degree of deterioration of the odor sensor 102.

In another aspect, when the odor sensor 102 of one odor detection device 100 deteriorates, the odor detection system 10 can complement the deteriorated odor sensor 102 of the odor detection device 100 by using the odor sensor 102 of another odor detection device 100.

The embodiments disclosed herein should be considered illustrative in all respects and not restrictive. The scope of the present disclosure is indicated by the claims, not by the above description, and is intended to include all modifications within the meaning and scope of the claims. Furthermore, the disclosures described in the embodiments and each modification are intended to be implemented, to the extent possible, either alone or in combination.

10 Odor detection system, 100 Odor detection device, 101, 121 Control unit, 102A, 102B, 102C, 102D Odor sensor, 104, 126 Memory unit, 105 Identification unit, 106 Odor type discrimination table, 107 Determination unit, 108 Correction unit, 109 Characteristic data, 110 Operation switch, 111 Battery, 112 Communication unit, 113 Internal bus, 120 Communication device, 122 Display unit, 123 Operation unit, 124 First communication unit, 125 Second communication unit, 127 Speaker, 128 Microphone, 301 Gas to be measured, 302 Record, 401, 402 Graph, 403 Threshold, 404 Regression curve.

Claims (12)

Three or more odor sensors;
a memory unit for storing normal range data obtained by learning the normal range of the output value of each odor sensor for each odor component;
a determination unit that determines the deterioration of each of the plurality of odor sensors,
The determination unit
comparing the output values of each of the plurality of odor sensors with the normal range data;
The odor detection device determines that the minority of the odor sensors have deteriorated when a majority of the plurality of odor sensors output actual measurement values within the normal range and the remaining minority of the odor sensors output actual measurement values outside the normal range . 2. The odor detection device of claim 1, wherein the determination unit determines that the odor components detected by the plurality of odor sensors are indistinguishable based on the fact that no odor components are found in the normal range data whose actual measurement values of the majority of odor sensors are within the normal range. The device further includes a correction unit that corrects the actual measurement value of the deteriorated odor sensor,
The correction unit
The actual measured value of a certain odor sensor is compared with an approximate line of the correlation between the concentration of each odor component learned in advance and the output value of the certain odor sensor ,
An odor detection device as described in claim 1 or 2, which corrects the actual measured value of a certain odor sensor to approach the output value indicated by the approximation line based on the difference between the actual measured value and the output value indicated by the approximation line being within a predetermined threshold value. The odor detection device of claim 3, wherein the determination unit outputs a determination result indicating that the odor sensor has deteriorated based on the difference between the actual measurement value of the odor sensor and the output value indicated by the approximation line exceeding a predetermined threshold. The odor detection device of claim 4, wherein the determination unit outputs an estimated time until the odor sensor becomes unusable from a regression curve obtained from time-dependent data of actual measured values of the odor sensor. The storage unit further stores recorded data of deterioration times of the odor sensor measured in the past,
The odor detection device according to claim 4 , wherein the determination unit compares the actual measurement value of the certain odor sensor with the recorded data, and outputs an estimated time until the certain odor sensor becomes unusable. An odor detection system, comprising:
A plurality of odor detection devices;
a communication device;
Each of the plurality of odor detection devices includes:
Three or more odor sensors;
a memory unit for storing normal range data obtained by learning the normal range of the output value of each odor sensor for each odor component;
a determination unit that determines the deterioration of each of the plurality of odor sensors,
The determination unit
comparing the output values of each of the plurality of odor sensors with the normal range data;
If a majority of the plurality of odor sensors output actual measurement values within a normal range and the remaining minority of odor sensors output actual measurement values outside the normal range, the minority of odor sensors are determined to be deteriorated;
The communication device
a communication unit for communicating with each of the plurality of odor detection devices;
an odor detection system comprising a display unit capable of displaying the results of deterioration determination obtained from each of the plurality of odor detection devices; 8. The odor detection system of claim 7, wherein the determination unit determines that the odor components detected by the plurality of odor sensors are indistinguishable based on the fact that no odor components are found in the normal range data whose actual measurement values of the majority of odor sensors are within the normal range. The odor detection device further includes a correction unit that corrects the actual measurement value of a deteriorated odor sensor,
The correction unit
The actual measured value of a certain odor sensor is compared with an approximate line of the correlation between the concentration of each odor component learned in advance and the output value of the certain odor sensor ,
The odor detection system of claim 7 or 8, wherein the actual measured value of the odor sensor is corrected to approach the output value indicated by the approximation line based on the difference between the actual measured value and the output value indicated by the approximation line being within a predetermined threshold. The odor detection system of claim 9, wherein the determination unit transmits to the communication device a determination result indicating that the certain odor sensor has deteriorated based on the difference between the actual measurement value of the certain odor sensor and the output value indicated by the approximation line exceeding a predetermined threshold. The odor detection system of claim 10, wherein the determination unit transmits to the communication device an estimated time until the odor sensor becomes unusable, based on a regression curve obtained from time-dependent data of actual measured values of the odor sensor. The storage unit further stores recorded data of deterioration times of the odor sensor measured in the past,
The odor detection system according to claim 10 , wherein the determination unit compares the actual measurement value of the certain odor sensor with the recorded data, and transmits to the communication device an estimated time until the certain odor sensor becomes unusable.