JP5126131B2 - Gas detector - Google Patents

Description

  The present invention relates to a gas detection device, and more particularly to a gas detection device that detects a detection target gas such as ethanol contained in human breath.

  Conventionally, in a gas sensor that detects a gas by utilizing a change in resistance value of a metal oxide semiconductor, data of a resistance value, sensor temperature, and time taken in a predetermined period is learned by a neural network, and the output of the gas is obtained by output of this network. There has been proposed a gas sensor that detects the type and gas concentration and issues an alarm based on the detection result (see, for example, Patent Document 1).

  The model value for each gas type is calculated and stored from the differential waveform of the sensor output value when the sensor heating temperature is changed, and the gas type is identified by comparing the measured value with the model value. A gas detection device that calculates a gas concentration from a sensor output value based on the identified gas type has been proposed (see, for example, Patent Document 2).

JP-A-6-130017 JP-A-1-311261

  However, in the techniques of Patent Document 1 and Patent Document 2, although the calculated gas concentration or the presence or absence of an alarm can be known, the accuracy of the calculated gas concentration is the calculated value. There is a problem that is unknown.

  The present invention has been made to solve the above problems, and it is an object of the present invention to provide a gas detection device that can not only obtain a value for estimating a gas concentration but also grasp the estimation accuracy. And

  In order to achieve the above object, a gas detection device according to a first aspect of the present invention includes a gas detection unit that detects a physical quantity related to the concentration of a detection target gas contained in a detection target gas and outputs a detection value; A response characteristic equation that represents a response characteristic of the gas detection unit and includes a coefficient indicating a final arrival value that the detection value finally reaches according to the response characteristic is used, and the value of the response characteristic expression is used as the detection value. And a calculation means for calculating a probability distribution expressed by the probability calculated for each of the coefficients having different values, and the probability distribution calculated by the calculation means is estimated as the probability distribution of the final arrival value. And a final arrival value probability distribution estimating means.

  According to the gas detection device of the first invention, the gas detection means detects a physical quantity related to the concentration of the detection target gas contained in the detection target gas and outputs a detection value, and the calculation means detects the gas detection Expressing the response characteristic of the means, and using the response characteristic equation that includes the coefficient that indicates the final value that the detection value finally reaches according to the response characteristic, the response characteristic value is expressed as the probability that the value will be the detection value. The probability distribution represented by the probability calculated for each coefficient having different values is calculated, and the final reaching value probability distribution estimating means estimates the probability distribution calculated by the calculating means as the probability distribution of the final reaching value.

  In this way, by estimating the probability distribution of the final arrival value of the detection value based on the response characteristic equation and detection value of the gas detection means, not only can the gas concentration be estimated using the final arrival value. Since the probability of the final arrival value is also known, the estimation accuracy can be grasped.

  Moreover, the gas detection apparatus of 1st invention respond | corresponds to the final arrival value or the expected value when the probability is the maximum, based on the probability distribution of the final arrival value estimated by the final arrival value probability distribution estimation means. A gas concentration estimating means for estimating a value obtained by multiplying the final reached value by a predetermined value as the concentration of the detection target gas can be configured. Thereby, the gas concentration can be estimated with high accuracy.

  In the gas detector of the first invention, the probability of the final arrival value exceeding a predetermined reference value in the probability distribution of the final arrival value estimated by the final arrival value probability distribution estimation means is a predetermined threshold value. In the above case, it can comprise including the determination means which determines with the said detection target gas being contained in the said detection target gas. In this way, in the probability distribution of the final arrival value, it is determined whether or not the detection target gas is included based on the probability exceeding the reference value, so that the determination reliability is improved.

  A gas detection device according to a second aspect of the present invention represents a gas detection means for detecting a physical quantity related to the concentration of a detection target gas contained in a detection target gas and outputting a detection value, and a response characteristic of the gas detection means. And a response characteristic equation including a coefficient indicating a final arrival value that the detection value finally reaches according to the response characteristic, and a value represented by the probability that the value of the response characteristic equation becomes the detection value, and a value Calculating means for calculating a probability distribution represented by the probability calculated for each of the different coefficients, and the concentration of the detection target gas obtained by multiplying the probability distribution calculated by the calculating means by a predetermined value. Gas concentration probability distribution estimation means for converting the probability distribution represented by the probability for each value indicated and estimating the converted probability distribution as the probability distribution of the concentration of the detection target gas.

  According to the gas detection device of the second invention, the gas detection means detects a physical quantity related to the concentration of the detection target gas contained in the detection target gas and outputs a detection value, and the calculation means detects the gas detection Expressing the response characteristic of the means, and using the response characteristic equation that includes the coefficient that indicates the final value that the detection value finally reaches according to the response characteristic, the response characteristic value is expressed as the probability that the value will be the detection value. And a probability distribution represented by the probability calculated for each coefficient having a different value, and the gas concentration probability distribution estimation means multiplies the probability distribution calculated by the calculation means by a predetermined value to the coefficient. It converts into the probability distribution represented by the probability for every value which shows the density | concentration of gas, and estimates the converted probability distribution as a probability distribution of the density | concentration of detection object gas.

  In this way, by estimating the probability distribution of the gas concentration that is easier to grasp based on the response characteristic equation and the detection value of the gas detection means, not only the gas concentration can be estimated, but also the gas concentration is obtained. Since the probability is also known, it is possible to grasp the estimation accuracy.

  Further, the gas detection device of the second invention is a value corresponding to a value or an expected value when the probability is maximum based on the probability distribution of the concentration of the detection target gas estimated by the gas concentration probability distribution estimation means. Can be configured to include a gas concentration estimating means for estimating the detection target gas concentration as the concentration of the detection target gas. Thereby, the gas concentration can be estimated with high accuracy.

  In the gas detector of the second invention, in the probability distribution of the concentration of the detection target gas estimated by the gas concentration probability distribution estimating means, the probability of the concentration exceeding a predetermined reference value is equal to or greater than a predetermined threshold value. In this case, it is possible to include a determination unit that determines that the detection target gas is contained in the detection target gas. As described above, in the gas concentration probability distribution, it is determined whether or not the detection target gas is included based on the probability exceeding the reference value, so that the determination reliability is improved.

  Moreover, the response characteristic formula of the gas detection means of the present invention can be expressed by the following formula.

  Here, a is a coefficient indicating the final reached value, and b and c are coefficients.

  As described above, according to the present invention, it is possible not only to obtain a value for estimating the gas concentration but also to estimate the probability distribution of the final arrival value used for estimating the gas concentration. The effect that accuracy can also be grasped is obtained.

It is the schematic which shows the state which attached the ethanol concentration detection apparatus of this Embodiment to the steering column of a driver's seat. It is the schematic which shows this Embodiment. It is a block diagram which shows schematic structure of the ethanol concentration detection apparatus of 1st Embodiment. It is a flowchart which shows the content of the gas detection process routine in the ethanol concentration detection apparatus of 1st Embodiment. It is a diagram which shows an example of probability distribution of the final arrival value of the detected value of an alcohol sensor. It is a block diagram which shows schematic structure of the ethanol concentration detection apparatus of 2nd Embodiment. It is a flowchart which shows the content of the gas detection process routine in the ethanol concentration detection apparatus of 2nd Embodiment. It is a diagram which shows an example of probability distribution of the gas concentration of ethanol gas.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the present embodiment, a case where the present invention is applied to an ethanol concentration detection device that detects the concentration of ethanol, which is a kind of alcohol, from the breath of a driver will be described as an example.

  As shown in FIG. 1, the ethanol concentration detection apparatus 10 according to the first embodiment is attached to a position where the driver's exhalation of the steering column 12 provided in the driver's seat can reach. The ethanol concentration detection device 10 includes an elongate cylindrical exhalation introduction tube 20 having a suction port 20 </ b> A having an enlarged diameter at the tip, and an alcohol sensor 24 is attached to the inside of the exhalation introduction tube 20. ing.

  As shown in FIG. 2, a suction fan 22 that is driven to suck the breath of the driver from the suction port 20A is provided inside the breath introduction tube 20 and closer to the suction port 20A than the alcohol sensor 24. .

  The alcohol sensor 24 is a sensor that detects ethanol gas contained in the gas flowing through the exhalation-introducing pipe 20, and for example, an oxide semiconductor gas sensor can be used. The alcohol sensor 24 outputs a detection signal having a higher level as the concentration of the ethanol gas contained in the gas flowing in the exhalation introduction tube 20 increases, and the ethanol sensor contained in the gas flowing in the exhalation introduction tube 20 A detection signal with a low level is output as the density decreases.

  According to the present embodiment, when the suction fan 22 is driven, the exhaled air exhaled from the driver is sucked into the exhalation introduction tube 20 from the suction port 20A of the exhalation introduction tube 20, and the exhalation is mixed with the air. As a result, the alcohol sensor 24 is diluted at a constant flow rate. The exhaled breath comes into contact with the alcohol sensor 24 and is then discharged out of the exhaled breath introducing tube 20.

  When the exhaled breath comes into contact with the alcohol sensor 24, a detection value related to the concentration of ethanol gas in the gas containing exhaled breath is detected by the alcohol sensor 24. Based on the detection value detected by the alcohol sensor 24, the concentration of ethanol gas as the detection target gas in expiration is estimated.

  As shown in FIG. 3, the ethanol concentration detection apparatus 10 includes an ethanol concentration detector 30 that is connected to an alcohol sensor 24, a display device 26 such as a liquid crystal display, and a speaker 28 and detects the concentration of ethanol gas. Yes.

  The ethanol concentration detector 30 includes a CPU that controls the entire ethanol concentration detector 10, a ROM as a storage medium that stores various programs such as a gas detection program described later, a RAM that temporarily stores data as a work area, and these Can be constituted by a microcomputer including a bus for connecting the two.

  The ethanol concentration detector 30 will be described in terms of functional blocks divided for each function realization means determined based on hardware and software. A detection value storage unit 40 for storing a detection value of the alcohol sensor 24, and a detection value storage unit Based on the detection value stored in 40 and the response characteristic formula of the alcohol sensor 24, a final arrival value probability distribution estimation unit 42 that estimates the probability distribution of the final arrival value of the detection value, and a final arrival value probability distribution estimation unit 42 The gas concentration estimator 46 that estimates the gas concentration of ethanol gas based on the final ultimate value that has the maximum probability in the probability distribution of the final ultimate value estimated in step S3, and the final estimated by the final ultimate value probability distribution estimation unit 42 Based on the probability distribution of the arrival value, the gas presence / absence determination unit 48 for determining whether or not ethanol gas exceeding the reference value is present in the expiration, and the estimation result of the gas concentration estimation unit 46 And a display control unit 50 that controls to display the determination result of the gas presence / absence determination unit 48 on the display device 26, and an output that controls to output an alarm from the speaker 28 based on the determination result of the gas presence / absence determination unit 48. It can be expressed by a configuration including the part 52.

  Next, the gas detection processing routine in the first embodiment will be described with reference to FIG. This routine is performed by the CPU executing a gas detection program stored in the ROM.

  In step 100, the heater of the alcohol sensor 24 is turned on, and acquisition of the detection value output from the alcohol sensor 24 is started.

  Next, in step 102, when exhalation continues to be introduced, the probability distribution of the value that the detection value finally reaches (final arrival value) is estimated. The specific method will be described below.

  As a response characteristic formula of the alcohol sensor 24, the detected value f (t) at time t when the time when the detected value of the alcohol sensor 24 starts to change is set to t = 0 using parameters a, b, and c. Modeling is performed as shown in the following equation (1).

  If t → ∞ (when exhalation continues to be introduced) in the above equation (1), the value of the detection value f (t) converges to the parameter a. That is, the parameter a is the final arrival value of the detected value, and the final arrival value can be estimated by estimating the parameters a, b, and c.

  Here, a case where the parameters a, b, and c are estimated using a particle filter that is a kind of Bayes filter will be described.

Particle i at time t = t (i = 1,2, ···, M: M is the number of particles (e.g., 1000)) the state ^x _{i t} of, using the parameters at state ^x _{i t} following (2) It represents like a type | formula.

And _{z t} the detection value of the alcohol sensor 24 at time t = t. In the state x ⁱ _t , the probability P (z _t | x ⁱ _t ) of observing the detected value z _t is the average f ⁱ _t and variance σ ₁ = [0.001, 0.25 of the following equation (3). , 0.05] normal distribution N (f ⁱ _t , σ ₁ ).

Furthermore, probability state ^x _{i t-1} after the state _{x t} is the observed _{^{P (x t | x i t}} -1) is the average ^x _{i t-1,} the dispersion σ ₂ = [0.002,0.05 , 0.01] according to the normal distribution N (x ⁱ _t−1 , σ ₂ ).

At this time, the probability distribution P (x _t | z _t ) of the state x ⁱ _t is estimated by calculating the following algorithm. However, the probability distribution P (x ₀ | z ₀ ) at time t = 0 is a uniform distribution, and particles at t = 0 have 0 <a ₀ <0.2, 0 <b ₀ <5.0, ₀ . It is assumed that the parameter space exists evenly in the range of <c ₀ <1.0.

  I. The following processing is performed for all M particles.

i One particle i is selected at random with a probability proportional to the probability distribution P (x _t | x ⁱ _t−1 ).

ii Calculate the weight w ⁱ _t = P (z _t | x ⁱ _t ) for particle i.

II. Duplication is permitted from among M particles, and M particles are selected with a probability proportional to the weight w ⁱ _t .

III. State ^x _{i t} of the probability distribution _{P (x t} | _z t) is approximated as the density of particles in the small section _{[x t -Δx, x t +} Δx].

By calculating the above algorithm at each time, the probability distribution P (x _t | z _t ) of the state x ⁱ _{t at} the current time is estimated. The probability distribution P (x _t | z _t ) of the state x ⁱ _t to be finally estimated is, for example, the probability distribution P (x _t | z of the state x ⁱ _{t at} the time when the detection value of the alcohol sensor 24 reaches the peak value. _t ).

Next, in step 104, based on the probability distribution P (x _t | z _t ) of the state x ⁱ _t estimated in step 102, the state x ^max _t having the maximum probability is selected. Each parameter in the state x ^max _t is _{expressed by} the following expression (4) from the expression (2), and the final reached value of the detected value is a ^max _t .

このａ^ｍａｘ _ｔを用いて下記（５）式により、アルコールセンサ２４の検出値の最終到達値をエタノールガスのガス濃度Ｅに変換する。

Using this a ^max _t , the final value of the detection value of the alcohol sensor 24 is converted into the gas concentration E of ethanol gas by the following equation (5).

  Here, ln is a natural logarithm, E is a gas concentration of ethanol gas, and k is a detection value concentration conversion coefficient calculated in advance.

Next, in step 106, the gas concentration E estimated in step 104 and the probability of the gas concentration are displayed on the display device 26. The probability of the gas concentration is the maximum probability used when the final reached value a ^max _t of the detected value is estimated in step 104 above. The estimated gas concentration E and the probability of the gas concentration may be output from the speaker 28 as a sound.

Next, in step 108, a reference value Th _a for determining whether or not the exhaled breath contains ethanol gas equal to or higher than the reference value in the probability distribution of the final arrival value of the detected value estimated in step 102 is set. calculating the area ratio P _a is exceeded. For example, in the case of the probability distribution of the final arrival value as shown in FIG. 5, the area ratio of the hatched portion in the figure in the entire area surrounded by the curve indicating the probability distribution is calculated. The reference value Th _a is determined in advance as a value obtained by converting the value of the ethanol concentration when it is determined that the exhaled gas contains ethanol gas into the detection value of the alcohol sensor 24 based on the above equation (5). .

Next, in step 110, it is determined whether or not the area ratio _Pa calculated in step 108 is equal to or greater than _a determination threshold value _Ptha . The determination threshold value P _tha is an index indicating how much the probability that the reference value Th _a is exceeded. For example, a value such as P _tha = 0.8 is determined in advance. When P _a ≧ P _tha, the process proceeds to step 112, an alarm such as a buzzer sound or a message is output from the speaker 28, and the process ends. If P _a <P _tha , the process ends. An alarm message may be displayed on the display device 26.

  As described above, according to the ethanol concentration detection apparatus of the first embodiment, in order to estimate the probability distribution of the final arrival value of the detection value of the alcohol sensor, the gas concentration estimated by converting the final arrival value is estimated. In addition to knowing the value of the gas, it is also possible to understand how accurately the gas concentration is estimated, presenting the estimation result of the gas concentration, and the gas using the estimation result The reliability of the presence / absence determination is improved.

  Next, a second embodiment will be described. In addition, about the structure and process same as 1st Embodiment, the same code | symbol is attached | subjected and description is abbreviate | omitted.

  In the ethanol concentration detection apparatus according to the first embodiment, the case of estimating the probability distribution of the final arrival value of the detection value of the alcohol sensor has been described. However, in the second embodiment, the final arrival value is converted into a gas concentration. However, it differs from the first embodiment in that the probability distribution of gas concentration is estimated.

  As shown in FIG. 6, an ethanol concentration estimation apparatus 210 according to the second embodiment is connected to an alcohol sensor 24, a display device 26, and a speaker 28, and detects an ethanol gas concentration. 230 is provided. The ethanol concentration detector 230 includes a gas concentration probability distribution estimation unit 44 that estimates a probability distribution of gas concentration based on the detection value stored in the detection value storage unit 40 and the response characteristic equation of the alcohol sensor 24, and a gas concentration. A gas concentration estimation unit 246 that estimates a gas concentration as a gas concentration probability value estimated by the probability distribution estimation unit 44, and a gas concentration estimated by the gas concentration probability distribution estimation unit 44. Based on the probability distribution, it can be expressed by a configuration including a gas presence / absence determination unit 248 that determines whether or not there is a gas equal to or higher than a reference value during expiration.

  Next, the gas detection processing routine in the second embodiment will be described with reference to FIG. This routine is performed by the CPU executing a gas detection program stored in the ROM.

  In step 100, the heater of the alcohol sensor 24 is turned on, and acquisition of the detection value output from the alcohol sensor 24 is started.

  Next, in step 300, the probability distribution of the gas concentration of ethanol gas contained in exhaled breath is estimated. The probability distribution of the final reached value estimated in the same manner as in the first embodiment is converted based on the above equation (5), and the converted probability distribution is estimated as the probability distribution of the gas concentration.

  Next, at step 302, based on the probability distribution of the gas concentration estimated at step 300, the value with the highest probability is estimated as the gas concentration E of ethanol gas. Next, in step 106, the gas concentration E estimated in step 302 and the probability of the gas concentration are displayed on the display device 26.

Next, in step 306, in the probability distribution of the gas concentration estimated in step 300 above, the area exceeding the reference value Th _E for determining whether or not the exhaled breath contains ethanol gas equal to or higher than the reference value The rate _PE is calculated. For example, in the case of the gas concentration probability distribution as shown in FIG. 8, the area ratio of the hatched portion in the figure in the entire area surrounded by the curve indicating the probability distribution is calculated. As the reference value Th _E , a value (for example, 0.15 mg) of ethanol concentration when it is determined that an ethanol component is contained in exhaled breath is determined in advance.

Next, in step 308, calculated in step 306 area ratio _{P E} determines whether the determination threshold _{P thE} more. The determination threshold value P _thE is an index indicating how much the probability of exceeding the reference value Th _E is, for example, a value such as P _thE = 0.8 is determined in advance. If P _E ≧ P _thE, the _routine proceeds to step 112, where an alarm such as a buzzer sound or a message is output from the speaker 28, and the processing is terminated. If P _E <P _thE , the process ends.

  As described above, according to the ethanol concentration detection apparatus of the second embodiment, in order to estimate the probability distribution of the gas concentration of ethanol gas, not only the gas concentration value can be known, but also the gas concentration is It can also be understood to what extent the value is estimated, and the reliability of the presentation of the gas concentration estimation result and the determination of the presence or absence of gas using the estimation result is improved.

  In the above embodiment, the particle filter is used when estimating the final distribution value or the probability distribution of the gas concentration. However, the present invention is not limited to this. For example, the probability distribution can be similarly estimated using an extended Kalman filter as described below.

In the system represented by the following equation (8), the state x _k is estimated.

Where x _k is the state at time k, u _k is the control input, w _k is mean zero, noise related to time transitions following the normal distribution of the covariance matrix Q _k , z _k is the observable, and v _k is mean zero, it is the observation noise follows a normal distribution of the covariance matrix R _k. At this time, the state is estimated by predicting and updating the value according to the following procedure.

ただし、

However,

  In this embodiment,

とすることにより、パラメータａ、ｂ、及びｃを推定することができる。

Thus, the parameters a, b, and c can be estimated.

  Further, in the above embodiment, the case where the gas concentration is estimated using the value that has the maximum probability in the final distribution value or the probability distribution of the gas concentration has been described. However, the gas concentration is estimated using the expected value. It may be.

  Further, in the above-described embodiment, the case where an alarm is given based on the presence or absence of ethanol gas in exhalation has been described. However, when it is determined that ethanol gas is contained in exhalation, the engine cannot be started. For example, control may be performed so as to prevent fraud.

  In the above embodiment, the case where ethanol is detected from the exhalation of the driver has been described. However, the present invention detects ethanol from the exhalation of humans other than the driver by configuring the ethanol concentration detection device to be portable. It can also be applied to cases.

DESCRIPTION OF SYMBOLS 10,210 Ethanol concentration detection apparatus 24 Alcohol sensor 26 Display apparatus 28 Speaker 30, 230 Ethanol concentration detector 40 Detection value storage part 42 Final arrival value probability distribution estimation part 44 Gas concentration probability distribution estimation part 46, 246 Gas concentration estimation part 48 248 Gas presence / absence determination unit 50 Display control unit 52 Output unit

Claims (7)

Gas detection means for detecting a physical quantity related to the concentration of the detection target gas contained in the detection target gas and outputting a detection value;
A response characteristic equation that represents a response characteristic of the gas detection unit and includes a coefficient indicating a final arrival value that the detection value finally reaches according to the response characteristic is used, and the value of the response characteristic expression is the detection value A calculation means for calculating a probability distribution expressed by the probability calculated for each of the coefficients having different values and represented by the probability of
A final reaching value probability distribution estimating means for estimating the probability distribution calculated by the calculating means as a probability distribution of the final reaching value;
A gas detection device.   A value obtained by multiplying the final arrival value corresponding to the final arrival value or the expected value when the probability is the maximum by a predetermined value based on the probability distribution of the final arrival value estimated by the final arrival value probability distribution estimation unit The gas detection apparatus according to claim 1, further comprising: a gas concentration estimation unit that estimates a gas concentration as a concentration of the detection target gas.   In the probability distribution of the final arrival value estimated by the final arrival value probability distribution estimation means, when the probability of the final arrival value exceeding a predetermined reference value is equal to or higher than a predetermined threshold, The gas detection apparatus according to claim 1, further comprising a determination unit that determines that the detection target gas is included. Gas detection means for detecting a physical quantity related to the concentration of the detection target gas contained in the detection target gas and outputting a detection value;
A response characteristic equation that represents a response characteristic of the gas detection unit and includes a coefficient indicating a final arrival value that the detection value finally reaches according to the response characteristic is used, and the value of the response characteristic expression is the detection value A calculation means for calculating a probability distribution expressed by the probability calculated for each of the coefficients having different values and represented by the probability of
The probability distribution calculated by the calculation means is converted into a probability distribution represented by a probability for each value indicating the concentration of the detection target gas obtained by multiplying the coefficient by a predetermined value, and the converted probability distribution is converted into the probability distribution. A gas concentration probability distribution estimating means for estimating the concentration distribution of the detection target gas as a probability distribution;
A gas detection device.   Based on the probability distribution of the concentration of the detection target gas estimated by the gas concentration probability distribution estimation means, a gas concentration that estimates a value corresponding to a maximum probability or a value corresponding to an expected value as the concentration of the detection target gas The gas detection apparatus according to claim 4, further comprising an estimation means.   In the probability distribution of the concentration of the detection target gas estimated by the gas concentration probability distribution estimation means, when the probability of concentration exceeding a predetermined reference value is equal to or higher than a predetermined threshold, the detection target gas contains the 6. The gas detection apparatus according to claim 4, further comprising a determination unit that determines that the detection target gas is contained. The gas detection device according to any one of claims 1 to 6, wherein a response characteristic formula of the gas detection means is expressed by the following formula.

Here, a is a coefficient indicating the final reached value, and b and c are coefficients.

Images