Optimization device and optimization method

Abstract

An optimization device (40) optimizes a sample time in data acquisition using a chemical sensor (20) that detects a measurement target object, the optimization device comprising: a feature amount extraction unit (42) that acquires waveform data indicated by a signal outputted from the chemical sensor (20), and extracts a feature amount included in the acquired waveform data; and a control unit (44) that determines, on the basis of the extracted feature amount, a sample time, which is the time during which the measurement target object is exposed to the chemical sensor (20), and outputs a signal indicating the determined sample time.

Description

Optimization Apparatus and Optimization Method

[0001] The present disclosure relates to an optimization apparatus and an optimization method, and particularly relates to an optimization apparatus for optimizing sample time in data acquisition using a chemical sensor that detects odors or the like.

[0002] Conventionally, various techniques for detecting odors using chemical sensors have been proposed (for example, Patent Documents 1 and 2, etc.).

[0003] Patent Document 1 discloses a technique for detecting an odor of a measurement target by using a reference odor, thereby stably detecting the odor with high accuracy regardless of temporal or environmental factors.

[0004] In addition, Patent Document 2 discloses a technique for accurately and efficiently collecting information related to odors by detecting odors both with and without a filter and collecting the difference between them as odor information.

[0005] International Publication No. 2022 / 085345; Japanese Patent No. 7235726

[0006] However, in any of the conventional techniques of Patent Documents 1 and 2, it is necessary to determine in advance the sample time, which is the time for which the measurement target is exposed to the sensor. However, the optimal value of the sample time may vary depending on the type of measurement target and other factors. For this reason, the conventional techniques have a problem that correct information cannot be obtained when the sample time is short, whereas data acquisition time becomes long when the sample time is long.

[0007] Accordingly, the present disclosure aims to provide an optimization apparatus and an optimization method capable of automatically determining an optimal sample time.

[0008] To achieve the above objective, an optimization device according to one embodiment of the present disclosure is an optimization device for optimizing the sample time in data acquisition using a chemical sensor, comprising: a feature extraction unit that acquires waveform data indicated by a signal output from a chemical sensor that detects a target for measurement, and extracts feature quantities contained in the acquired waveform data; and a control unit that determines the sample time, which is the time the target for measurement is exposed to the chemical sensor, based on the extracted feature quantities, and outputs a signal indicating the determined sample time.

[0009] To achieve the above objective, an optimization method according to one embodiment of the present disclosure is an optimization method for optimizing the sample time in data acquisition using a chemical sensor, comprising: a feature extraction step of acquiring waveform data indicated by a signal output from a chemical sensor that detects a target for measurement, and extracting feature quantities contained in the acquired waveform data; and a control step of determining a sample time, which is the time the target for measurement is exposed to the chemical sensor, based on the extracted feature quantities, and outputting a signal indicating the determined sample time.

[0010] This disclosure provides an optimization device and an optimization method that can automatically determine the optimal sample time.

[0011] Figure 1 is a block diagram showing the configuration of a sensor system according to an embodiment. Figure 2 is a flowchart showing the operation of the sensor system according to an embodiment. Figure 3A is a flowchart showing one specific method for determining the sample time by the optimization device according to an embodiment. Figure 3B is a diagram illustrating the determination method shown in the flowchart of Figure 3A. Figure 4A is a flowchart showing yet another specific method for determining the sample time by the optimization device according to an embodiment. Figure 4B is a diagram illustrating a specific method for calculating the similarity shown in the flowchart of Figure 4A. Figure 4C is a diagram illustrating the determination method shown in the flowchart of Figure 4A. Figure 5A is a flowchart showing yet another specific method for determining the sample time by the optimization device according to an embodiment. Figure 5B is a diagram illustrating the determination method shown in the flowchart of Figure 5A. Figure 6A is a flowchart showing one specific method for determining a common sample time for multiple sensor elements by the optimization device according to an embodiment. Figure 6B is a diagram illustrating the determination method shown in the flowchart of Figure 6A. Figure 7A is a flowchart showing yet another specific method for determining a common sample time for multiple sensor elements by the optimization device according to an embodiment. Figure 7B is a diagram illustrating the determination method shown in the flowchart of Figure 7A.

[0012] The embodiments of this disclosure will be described in detail below with reference to the drawings. The embodiments described below are all specific examples of this disclosure. The numerical values, materials, components, arrangement and connection configurations of components, steps, step order, signal waveforms, feature quantities, etc., shown in the following embodiments are examples only and are not intended to limit this disclosure. Furthermore, the figures are not necessarily strictly illustrative. In each figure, substantially identical components are denoted by the same reference numerals, and redundant explanations are omitted or simplified.

[0013] Figure 1 is a block diagram showing the configuration of a sensor system 50 according to an embodiment. The sensor system 50 is a system that detects the odor of a target object using a chemical sensor, and consists of a chemical sensor 20, a waveform acquisition device 30, and an optimization device 40.

[0014] The chemical sensor 20 is a sensor that responds specifically or selectively to a specific chemical substance, which is the target of measurement, and outputs a signal of a magnitude proportional to its amount. Examples include odor sensors, gas sensors, ion sensors, biosensors, etc. In this embodiment, a plurality of sensor elements 1 to 16, made of different materials and having different response characteristics, are arranged in a two-dimensional manner. The target of measurement may be a sample with unknown characteristics, a reference with known characteristics, or both a sample and a reference.

[0015] The waveform acquisition device 30 is a device that acquires signals output from each of the multiple sensor elements 1 to 16 that constitute the chemical sensor 20 and generates waveform data from the acquired signals. For example, it is a resistance measuring instrument that measures the resistance value of each of the multiple sensor elements 1 to 16 and outputs the voltage representing the measured resistance value as waveform data, either as an analog signal or as a time series of digital values ​​obtained by A / D conversion.

[0016] The optimization device 40 is a device that optimizes the sample time in data acquisition using the chemical sensor 20. For example, it is implemented as a computer device and functionally has a feature extraction unit 42 and a control unit 44. The feature extraction unit 42 and the control unit 44 are implemented, for example, by a processor that executes a program.

[0017] The feature extraction unit 42 acquires waveform data output from the waveform acquisition device 30 for each of the multiple sensor elements 1 to 16, and extracts feature quantities contained in the acquired waveform data. These feature quantities include, for example, (1) the slope, which is the time change in the magnitude of the signal shown by the waveform data; (2) the similarity between the waveform shown by the waveform data and a pre-held reference waveform; and (3) the ratio of the signal magnitude, which is the upper limit of the signal shown by the waveform data, to a predetermined upper limit. Which feature quantities to adopt is determined by prior settings by the user.

[0018] The control unit 44 determines the sample time, which is the time the object to be measured is exposed to the chemical sensor 20, based on the features extracted by the feature extraction unit 42, and outputs a signal indicating the determined sample time. At this time, the control unit 44 may (1) determine the timing when all of the feature quantities of the plurality of waveform data corresponding to each of the plurality of sensor elements 1 to 16 satisfy a predetermined first condition, and determine a sample time ending at the timing when the predetermined first condition is satisfied, or (2) determine the timing when any of the feature quantities of the plurality of waveform data corresponding to each of the plurality of sensor elements 1 to 16 satisfy a predetermined second condition, and determine a sample time ending at the timing when the predetermined second condition is satisfied.

[0019] Furthermore, the control unit 44 has a distinctive method for determining the end point of the sample time. The start point of the sample time may be the timing when the object to be measured is first exposed to the chemical sensor 20, or the timing when a signal indicating this is received, and it is not even necessary to determine the start point of the sample time.

[0020] Furthermore, the control unit 44 may perform qualitative and quantitative analysis of the measurement target, such as odor discrimination, from the features of the waveform data extracted by the feature extraction unit 42 during the sample time.

[0021] Figure 2 is a flowchart showing the operation of the sensor system 50 according to the embodiment (i.e., the operation including the optimization method by the optimization device 40). First, the waveform acquisition device 30 acquires signals output from each of the multiple sensor elements 1 to 16 that constitute the chemical sensor 20, and generates waveform data indicated by the acquired signals (S10).

[0022] Next, the feature extraction unit 42 of the optimization device 40 acquires waveform data output from the waveform acquisition device 30 for each of the multiple sensor elements 1 to 16, and extracts feature quantities contained in the acquired waveform data (feature extraction step; S11).

[0023] Next, the control unit 44 determines the sample time, which is the time the object to be measured is exposed to the chemical sensor 20, based on the features extracted by the feature extraction unit 42 (control step; S12), and outputs a signal indicating the determined sample time (control step; S13). At this time, the control unit 44 either (1) determines the timing when all of the feature quantities of the multiple waveform data corresponding to each of the multiple sensor elements 1 to 16 satisfy a predetermined first condition, and determines a sample time ending at the timing when the predetermined first condition is satisfied, or (2) determines the timing when any of the feature quantities of the multiple waveform data satisfy a predetermined second condition, and determines a sample time ending at the timing when the predetermined second condition is satisfied.

[0024] Figure 3A is a flowchart illustrating one specific method for determining the sample time using the optimization device 40 according to the embodiment. In other words, it shows a specific example of steps S11 and S12 shown in Figure 2. Figure 3B is a diagram illustrating the determination method shown in the flowchart of Figure 3A.

[0025] As shown in Figure 3A, in this determination method, the feature extraction unit 42 extracts the slope, which is the time change in the magnitude of the signal shown by the waveform data acquired from the waveform acquisition device 30, as a feature for each of the multiple sensor elements 1 to 16 (S20). Then, the control unit 44 determines whether or not the magnitude shown by the signal has saturated based on the slope extracted by the feature extraction unit 42 (S21). If it is determined that it has not saturated (No in S21), the determination is repeated, and if it is determined that it has saturated (Yes in S21), a sample time is determined with the timing of the determination as the endpoint (S22).

[0026] Specifically, let's assume that, based on user settings, the criterion for determining whether the slope has saturated is, for example, that the state where the slope is below a threshold of 0.1 continues for a predetermined period of time. Then, as shown in Figure 3B, the control unit 44 calculates the slope of the acquired waveform data. For example, if the calculated slope is 0.5, it continues the sample time (Figure 3B(a)). After that, when the slope becomes 0.09 and falls below the threshold of 0.1, it starts monitoring whether this state continues for a predetermined period of time (Figure 3B(b)). If the state where the slope is below the threshold of 0.1 continues for a predetermined period of time, it determines that the waveform data has saturated, determines the sample time (in this example, the measurement time for the sample) with that timing as the endpoint, ends the measurement, and controls the system to switch the measurement target to the reference (Figure 3B(c)). In this way, the optimal sample time is automatically determined based on the slope of the waveform data. Note that the threshold can be set arbitrarily. The same applies to other cases.

[0027] Switching the measurement target from the sample to the reference is not always necessary, but it is done, for example, when measuring the reference after the sample, or when purging the previously measured sample from the space exposed to the chemical sensor 20 in order to measure another sample. Furthermore, the sample time for the reference after switching may be automatically determined using the same method as for the sample.

[0028] Figure 4A is a flowchart illustrating another specific method for determining the sample time by the optimization device 40 according to the embodiment. In other words, it shows another specific example of steps S11 and S12 shown in Figure 2. Figure 4B is a diagram illustrating the specific method for calculating the similarity shown in the flowchart of Figure 4A. Figure 4C is a diagram illustrating the determination method shown in the flowchart of Figure 4A.

[0029] As shown in Figure 4A, in this determination method, the feature extraction unit 42 extracts the similarity (here, the DTW (Dynamic Time Warping) distance) between the waveform data acquired from the waveform acquisition device 30 and a pre-held reference waveform for each of the multiple sensor elements 1 to 16 as a feature (S30). The control unit 44 then determines whether the state in which the similarity extracted by the feature extraction unit 42 was high continued for a certain period of time (S31). If it is determined that the state in which the similarity was high did not continue for a certain period of time (No in S31), the determination is repeated. If it is determined that the state in which the similarity was high continued for a certain period of time (Yes in S31), the control unit 44 determines a sample time with the determined timing as the endpoint (S32).

[0030] In this embodiment, the similarity between the waveform data and the reference waveform is measured using, for example, the DTW distance. As shown in Figure 4B, the DTW distance is calculated by exhaustively determining the distance between each point in the time series of the two waveform data sets (here, the first waveform data and the second waveform data), i.e., the absolute value of the error, and then finding the shortest path between the two time series. This similarity can be calculated even if the lengths and periods of the time series are different. It is an effective similarity measure when the periods are different but the shapes are similar, and a smaller number indicates a higher degree of similarity.

[0031] A specific example of the determination method shown in Figure 4A will be explained using Figure 4C. Let's assume that, as set by the user, the criterion for determining the timing of the endpoint is set to a predetermined period of time during which the similarity between the waveform shown in the waveform data and the reference waveform remains below a threshold of 1.0. Then, as shown in Figure 4C, the control unit 44 calculates the similarity between the waveform shown in the acquired waveform data and the reference waveform. If the calculated similarity is, for example, 5.0, the sample time is continued (Figure 4C(a)). Subsequently, when the similarity becomes 0.09 and falls below the threshold of 1.0, the control unit 44 begins monitoring whether this state continues for a predetermined period of time (Figure 4C(b)). If the state of similarity being below the threshold of 1.0 continues for a predetermined period of time, the control unit 44 determines that the waveform shown in the waveform data and the reference waveform are similar, determines the sample time (in this example, the measurement time for the sample) with that timing as the endpoint, ends the measurement, and controls the system to switch the measurement target to the reference (Figure 4C(c)). In this way, the optimal sample time is automatically determined based on the similarity between the waveform data and the reference waveform.

[0032] Figure 5A is a flowchart illustrating yet another specific method for determining the sample time by the optimization device 40 according to the embodiment. In other words, it shows a specific example of steps S11 and S12 shown in Figure 2. Figure 5B is a diagram illustrating the determination method shown in the flowchart of Figure 5A.

[0033] As shown in Figure 5A, in this determination method, the feature extraction unit 42 extracts as a feature the ratio of the signal magnitude, which is the magnitude of the signal shown by the waveform data acquired from the waveform acquisition device 30, to a predetermined upper limit for each of the multiple sensor elements 1 to 16 (S40). The control unit 44 then determines whether the ratio extracted by the feature extraction unit 42 exceeds a predetermined constant percentage (S41). If it determines that the ratio does not exceed the predetermined constant percentage (No in S41), the determination is repeated. If it determines that the ratio exceeds the predetermined constant percentage (Yes in S41), it recognizes that the magnitude of the signal shown by the waveform data has saturated and determines a sample time with the determined timing as the endpoint (S42).

[0034] Specifically, let's assume that, based on user settings, the criterion for determining whether the signal has saturated is, for example, when the signal exceeds 95% of a predetermined upper limit. Then, as shown in Figure 5B, the control unit 44 calculates the signal amount of the acquired waveform data. If the calculated signal amount is 80% of the upper limit, it continues the sampling time (Figure 5B(a)). Subsequently, if the signal amount is 90% of the upper limit, it also continues the sampling time (Figure 5B(b)). Furthermore, when the signal amount is 96% of the upper limit, it recognizes that the waveform data has saturated because it has exceeded 95% of the upper limit. It then determines a sample time (in this example, the measurement time for the sample) with that timing as the endpoint, ends the measurement, and controls the system to switch the measurement target to the reference (Figure 5B(c)). In this way, the optimal sampling time is automatically determined based on the signal amount of the waveform data.

[0035] Figure 6A is a flowchart illustrating one specific method for determining a common sample time for multiple sensor elements 1 to 16 using the optimization device 40 according to the embodiment. In other words, it shows a specific example of step S12 shown in Figure 2. Here, multiple sensor elements are also referred to as multiple "channels". Figure 6B is a diagram illustrating the determination method shown in the flowchart of Figure 6A.

[0036] As shown in Figure 6A, in this determination method, the control unit 44 determines whether each of the features of the multiple waveform data (i.e., channels) extracted by the feature extraction unit 42 satisfies a predetermined first condition (S50). If it is determined that any of the features does not satisfy the predetermined first condition (No in S50), the determination is repeated. If it is determined that all of the features satisfy the predetermined first condition (Yes in S50), the control unit determines a sample time ending at the timing when the predetermined first condition is satisfied (S51). Here, the predetermined first condition is determined by the user's prior settings and is at least one selected from the conditions shown in the flowchart of Figure 3A (S20-S22), the conditions shown in the flowchart of Figure 4A (S30-S32), and the conditions shown in the flowchart of Figure 5A (S40-S42).

[0037] Specifically, let's assume that, as set by the user, a predetermined first condition is set, for example, that the slope of the waveform data remains below a threshold of 0.1 for a predetermined period of time, and that the similarity between the waveform data and the reference waveform remains below a threshold of 1.0 for a predetermined period of time. Then, if there is a channel in which the feature quantities do not satisfy the predetermined first condition, the control unit 44 continues the sample time (Figure 6B(a)), and then, when the feature quantities of all channels satisfy the predetermined first condition, it determines the sample time (in this example, the measurement time for the sample) with that timing as the endpoint, ends the measurement, and controls the system to switch the measurement target to the reference (Figure 6B(b)). In this way, the optimal sample time with the endpoint being when the feature quantities of all channels satisfy the predetermined first condition is automatically determined.

[0038] Figure 7A is a flowchart illustrating another specific method for determining a common sample time for multiple sensor elements 1 to 16 using the optimization device 40 according to the embodiment. In other words, it shows a specific example of step S12 shown in Figure 2. Here, multiple sensor elements are also referred to as multiple "channels". Figure 7B is a diagram illustrating the determination method shown in the flowchart of Figure 7A.

[0039] As shown in Figure 7A, in this determination method, the control unit 44 determines whether any of the features of the multiple waveform data (i.e., channels) extracted by the feature extraction unit 42 satisfy a predetermined second condition (S60). If it is determined that none of the features satisfy the predetermined second condition (No in S60), the determination is repeated. If it is determined that any of the features satisfy the predetermined second condition (Yes in S60), the control unit determines a sample time ending at the timing when the predetermined second condition is satisfied (S61). Here, the predetermined second condition is determined by prior user settings and is at least one selected from the conditions shown in the flowchart of Figure 3A (S20-S22), the conditions shown in the flowchart of Figure 4A (S30-S32), and the conditions shown in the flowchart of Figure 5A (S40-S42).

[0040] Specifically, let's assume that, as set by the user, a predetermined second condition is set, for example, that the ratio of the waveform data to the upper limit of the signal amount (i.e., the feature quantity) exceeds 95% (i.e., an NG (No Good) judgment is issued). Then, if the feature quantity of any channel does not satisfy the predetermined second condition (i.e., no NG judgment is issued), the control unit 44 continues the sampling time (Figure 7B(a)). After that, when the feature quantity of any channel satisfies the predetermined second condition (i.e., a channel resulting in an NG judgment is generated), the control unit 44 determines the sampling time (in this example, the measurement time for the sample) with that timing as the endpoint, ends the measurement, and controls the system to switch the measurement target to the reference (Figure 7B(b)). In this way, the optimal sampling time with the endpoint being when the feature quantity of any channel satisfies the predetermined second condition is automatically determined.

[0041] The following technologies are disclosed based on the above description of embodiments.

[0042] (Technology 1) An optimization apparatus 40 for optimizing a sampling time in data acquisition using a chemical sensor 20, comprising: a feature extraction unit 42 that acquires waveform data indicated by a signal output from the chemical sensor 20 that detects a measurement object, and extracts a feature amount included in the acquired waveform data; and a control unit 44 that determines the sampling time, which is the time for which the measurement object is exposed to the chemical sensor 20, based on the extracted feature amount, and outputs a signal indicating the determined sampling time.

[0043] With this configuration, the sampling time is automatically determined from the feature amount of the waveform data, and accurate detection using the chemical sensor 20 is ensured without extra labor.

[0044] (Technology 2) The optimization apparatus 40 according to Technology 1, wherein the feature extraction unit 42 extracts, as the feature amount, a gradient which is a temporal change in the magnitude of the signal indicated by the waveform data, and the control unit 44 determines timing at which the magnitude indicated by the signal saturates based on the gradient, and determines the sampling time with the saturation timing as the end point. With this configuration, an appropriate sampling time that ends when the magnitude of the signal indicated by the waveform data saturates is determined.

[0045] (Technology 3) The optimization apparatus 40 according to any one of Technologies 1 to 2, wherein the feature extraction unit 42 extracts, as the feature amount, a similarity between a waveform indicated by the waveform data and a reference waveform stored in advance, and the control unit 44 determines timing at which the waveform becomes similar to the reference waveform based on the similarity, and determines the sampling time with the similar timing as the end point. With this configuration, an appropriate sampling time that ends when the waveform indicated by the waveform data becomes similar to the reference waveform is determined.

[0046] (Technology 4) The optimization apparatus 40 according to any one of Technologies 1 to 3, wherein the feature extraction unit 42 extracts, as the feature amount, a ratio of a signal amount, which is the magnitude of the signal indicated by the waveform data, to a predetermined upper limit, and the control unit 44 determines timing at which the signal amount saturates based on the ratio, and determines the sampling time with the saturation timing as the end point. With this configuration, an appropriate sampling time that ends when the magnitude of the signal indicated by the waveform data reaches a predetermined ratio of the upper limit is determined.

[0047] (Technology 5) The optimization apparatus 40 according to any one of Technologies 1 to 4, wherein the measurement object is a reference when a reference with known characteristics is exposed to the chemical sensor 20. With this configuration, the optimal sampling time for the reference is automatically determined.

[0048] (Technology 6) The optimization apparatus 40 according to any one of Technologies 1 to 5, wherein the chemical sensor 20 is composed of a plurality of sensor elements 1 to 16 that detect a measurement object, the feature amount extraction unit 42 acquires a plurality of waveform data indicated by signals output from the plurality of sensor elements 1 to 16, extracts feature amounts included in each of the plurality of acquired waveform data, and the control unit 44 determines a sampling time based on each of the feature amounts of the plurality of extracted waveform data, and outputs a signal indicating the determined sampling time. With this configuration, an appropriate sampling time common to the plurality of sensor elements 1 to 16 is determined in consideration of the entirety of the plurality of sensor elements 1 to 16.

[0049] (Technology 7) The optimization apparatus 40 according to Technology 6, wherein the control unit 44 determines a timing at which each of the feature amounts of the plurality of extracted waveform data all satisfy a predetermined first condition, and determines a sampling time with the timing at which the predetermined first condition is satisfied as an end point. With this configuration, an appropriate sampling time common to the plurality of sensor elements 1 to 16 is determined, with the end point being a point in time when all of the waveform data from the plurality of sensor elements 1 to 16 satisfy the predetermined first condition.

[0050] (Technology 8) The optimization apparatus 40 according to Technology 6, wherein the control unit 44 determines a timing at which any one of the feature amounts of the plurality of extracted waveform data satisfies a predetermined second condition, and determines a sampling time with the timing at which the predetermined second condition is satisfied as an end point. With this configuration, an appropriate sampling time common to the plurality of sensor elements 1 to 16 is determined, with the end point being a point in time when any one of the waveform data from the plurality of sensor elements 1 to 16 satisfies the predetermined second condition.

[0051] (Technical 9) An optimization method for optimizing the sample time in data acquisition using a chemical sensor 20, comprising: a feature extraction step of acquiring waveform data indicated by a signal output from a chemical sensor 20 that detects a target to be measured, and extracting feature quantities contained in the acquired waveform data; and a control step of determining the sample time, which is the time the target to be measured is exposed to the chemical sensor 20, based on the extracted feature quantities, and outputting a signal indicating the determined sample time.

[0052] With this configuration, the sample time is automatically determined from the features of the waveform data, ensuring accurate detection using the chemical sensor 20 without any extra effort.

[0053] (Technical 10) A program that causes a computer to execute the steps included in the optimization method described in Technical 9. With this configuration, the sample time is automatically determined from the feature quantities of the waveform data, and accurate detection using the chemical sensor 20 is ensured without any extra effort.

[0054] The optimization apparatus and optimization method relating to this disclosure have been described above based on embodiments, but this disclosure is not limited to these embodiments. Within the scope of this disclosure, various modifications to these embodiments that a person skilled in the art could conceive, as long as they do not depart from the spirit of this disclosure, and other forms constructed by combining some of the components of the embodiments, are also included.

[0055] For example, in the sensor system 50 according to the embodiment, the chemical sensor 20 is configured by arranging a plurality of sensor elements 1 to 16 having different response characteristics in a two-dimensional manner. However, the system is not limited to this configuration, and may consist of a single sensor element or a plurality of sensor elements arranged in a one-dimensional manner.

[0056] Furthermore, the optimization device 40 according to this embodiment is not limited to application to a sensor system for detecting odors, but can also be applied to a sensor system that uses a chemical sensor 20 to perform qualitative and quantitative analysis of gases or liquids.

[0057] Furthermore, in determining the end point of the sample time shown in Figure 3B, the end point of the sample time was determined to be the moment when the state in which the slope was below a predetermined threshold continued for a predetermined period of time. However, this predetermined period of time is not necessarily required and may be zero or greater.

[0058] Similarly, in determining the end point of the sample time shown in Figure 4C, the end point of the sample time was determined to be the timing when the similarity between the waveform shown by the waveform data and the reference waveform remained above a predetermined threshold for a predetermined period of time. However, this predetermined period of time is not necessarily required and may be zero or greater.

[0059] Furthermore, in this embodiment, the feature extraction unit 42 and the control unit 44 were implemented by a processor that executes a program, but the implementation is not limited to this form and may be implemented by hardware consisting of electronic circuits that perform signal processing.

[0060] Furthermore, a program that causes a computer to execute the steps included in the optimization method relating to this disclosure may be distributed as a non-temporary computer-readable recording medium such as a DVD, or as a program product.

[0061] Furthermore, the optimization apparatus according to this disclosure may also be an apparatus comprising an optimization apparatus 40 and a waveform acquisition apparatus 30. Moreover, the invention according to this disclosure can be realized not only as an optimization apparatus and optimization method, but also as a sensor system 50 consisting of a chemical sensor 20, a waveform acquisition apparatus 30, and an optimization apparatus 40, as shown in Figure 1.

[0062] This disclosure can be used as an optimization device for optimizing the sample time in data acquisition using chemical sensors that detect odors, etc., and in particular as an optimization device in a measurement system that can automatically determine the optimal sample time.

[0063] 1-16 Sensor elements 20 Chemical sensors 30 Waveform acquisition device 40 Optimization device 42 Feature extraction unit 44 Control unit 50 Sensor system

Claims

1. An optimization device for optimizing the sample time in data acquisition using a chemical sensor, comprising: a feature extraction unit that acquires waveform data indicated by a signal output from a chemical sensor that detects a target for measurement, and extracts feature quantities contained in the acquired waveform data; and a control unit that determines the sample time, which is the time the target for measurement is exposed to the chemical sensor, based on the extracted feature quantities, and outputs a signal indicating the determined sample time.

2. The optimization apparatus according to claim 1, wherein the feature extraction unit extracts the slope, which is the time change in the magnitude of the signal shown in the waveform data, as the feature quantity, and the control unit determines the timing at which the magnitude shown in the signal saturates based on the slope, and determines a sample time with the saturation timing as the endpoint.

3. The optimization apparatus according to claim 1, wherein the feature extraction unit extracts the similarity between the waveform shown in the waveform data and a pre-held reference waveform as the feature quantity, and the control unit determines the timing at which the waveform is similar to the reference waveform based on the similarity, and determines a sample time with the similar timing as the endpoint.

4. The optimization apparatus according to claim 1, wherein the feature extraction unit extracts a feature quantity as the ratio of the signal quantity, which is the magnitude of the signal shown by the waveform data, to a predetermined upper limit, and the control unit determines the timing at which the signal quantity saturates based on the ratio, and determines a sample time with the saturation timing as the endpoint.

5. The optimization apparatus according to claim 1, wherein the object to be measured is the reference in the case where a reference with known characteristics is exposed to the chemical sensor.

6. The optimization apparatus according to claim 1, wherein the chemical sensor is composed of a plurality of sensor elements that detect the object to be measured, the feature extraction unit acquires a plurality of waveform data indicated by signals output from the plurality of sensor elements, extracts feature quantities contained in each of the acquired plurality of waveform data, and the control unit determines the sample time based on the feature quantities of each of the extracted plurality of waveform data, and outputs a signal indicating the determined sample time.

7. The optimization apparatus according to claim 6, wherein the control unit determines the timing at which each of the feature quantities of the extracted plurality of waveform data satisfies a predetermined first condition, and determines a sample time with the timing at which the predetermined first condition is satisfied as the endpoint.

8. The optimization apparatus according to claim 6, wherein the control unit determines the timing at which any of the feature quantities of the extracted plurality of waveform data satisfy a predetermined second condition, and determines a sample time ending at the timing at which the predetermined second condition is satisfied.

9. An optimization method for optimizing the sample time in data acquisition using a chemical sensor, comprising: a feature extraction step of acquiring waveform data indicated by a signal output from a chemical sensor that detects a target for measurement, and extracting feature quantities contained in the acquired waveform data; and a control step of determining a sample time, which is the time the target for measurement is exposed to the chemical sensor, based on the extracted feature quantities, and outputting a signal indicating the determined sample time.

10. A program that causes a computer to perform the steps included in the optimization method described in claim 9.

Images