PORTABLE DEVICE FOR ESTIMATING THE PARTIAL PRESSURE OF BLOOD GAS

Abstract

A measuring device, intended to be placed on the skin of a human-being or animal user, the device extending between a contact face and a far end, the device including a side wall that extends between the contact face and the far end, on the contact face, at least one aperture to collect a gas of interest emitted through the skin, the aperture being produced through the contact face; a collecting chamber that extends between the contact face and a measuring chamber; a gas sensor, placed in the measuring chamber, to measure a concentration of gas of interest flowing through the measuring chamber; a processing unit, to estimate a gas-of-interest content in the body of the user based on the measurement taken by the gas sensor; a heat source to bring the contact face to a temperature above 37° C.; and an air intake produced through the side wall to admit ambient air into the duct.

Claims

1-20. (canceled)

21. A measuring device, configured to be placed on the skin of a user, the user being a human being or an animal, the device extending between a contact face, configured to be applied against the skin of the user, and a far end, the device comprising a side wall, that extends between the contact face and the far end, the device comprising: on the contact face, at least one aperture, configured to collect a transcutaneous gas of interest emitted through the skin, the aperture being formed through the contact face; a collecting chamber, extending between the contact face and a measuring chamber; a gas sensor, placed in the measuring chamber, and configured to measure a concentration of the gas of interest flowing through the measuring chamber; a processing and computing unit, configured to estimate a gas-of-interest content in the body of the user based on the measurement provided by the gas sensor; and a heat source, configured to bring the contact face to a temperature above 37° C., wherein the device further comprises: an aperture, forming an air intake, the aperture being formed through the side wall and configured to admit ambient air into the collecting chamber such that, under an effect of the temperature produced by the heat source, a convection current of air is generated, so that the air admitted into the collecting chamber through the side wall carrying the transcutaneous gas of interest to the measuring chamber, the air and the transcutaneous gas of interest forming a gas mixture extending into the measuring chamber.

22. The device of claim 21, wherein the gas of interest is carbon dioxide.

23. The device of claim 21, comprising a humidity sensor for determining a humidity level in the measuring chamber, wherein the processing and computing unit is configured to take into account the humidity level in the measuring chamber to estimate a concentration of gas of interest in the measuring chamber.

24. The device of claim 21, comprising a pressure sensor for determining a pressure in the measuring chamber, wherein the processing and computing unit is configured to take into account the pressure in the measuring chamber to estimate a partial pressure of gas of interest in the collecting chamber.

25. The device of claim 21, wherein the gas sensor is an optical sensor that comprises an infrared radiation source and a photodetector, the infrared radiation source and the photodetector being arranged such that the gas mixture flows, in the measuring chamber, between the infrared radiation source and the photodetector, the photodetector comprising two measurement channels configured to measure infrared radiation, emitted by the infrared radiation source, and having propagated through the measuring chamber in respectively: a measurement spectral band that correspond to an absorption spectral band of the gas of interest; and a reference spectral band, considered as not absorbed by the gas of interest.

26. The device of claim 21, further comprising: an auxiliary measuring chamber, configured to be filled by the ambient air extending around the device; and an auxiliary gas-of-interest sensor, placed in the auxiliary measuring chamber, and configured to measure a concentration of gas of interest in the ambient air filling the auxiliary measuring chamber, wherein the processing and computing unit is configured to take into account the measurement of the auxiliary sensor to estimate the gas-of-interest content in the body of the user.

27. The device of claim 26, wherein the auxiliary gas-of-interest sensor is an optical sensor that comprises an auxiliary infrared radiation source and an auxiliary photodetector, the auxiliary infrared radiation source and the auxiliary photodetector being arranged such that ambient air extends, in the measuring chamber, between the auxiliary infrared radiation source and the auxiliary photodetector, the auxiliary photodetector comprising two measurement channels configured to measure infrared radiation, emitted by the auxiliary infrared radiation source, and having propagated through the auxiliary measuring chamber in, respectively: a measurement spectral band that correspond to an absorption spectral band of the gas of interest; and a reference spectral band, considered as not absorbed by the gas of interest.

28. The device of claim 21, comprising: a sensor of ambient air pressure, the ambient air extending around the device; and/or a sensor of ambient air temperature; and/or a sensor of pressure in the collecting chamber; and/or a sensor of temperature in the collecting chamber; and/or a temperature sensor, configured to measure a skin temperature; and/or a sensor of temperature in the measuring chamber.

29. A method for estimating a gas-of-interest content in the blood of a user, using the device of claim 21, the device being applied such that the contact face is placed facing the skin of the user, the method comprising heating the contact face and admitting ambient air through the side wall of the device, the method comprising: a) estimating a concentration of gas of interest in the measuring chamber; b) based on the concentration of gas of interest resulting from a), computing a partial pressure of gas of interest in the collecting chamber; and c) based on the partial pressure of gas of interest in the collecting chamber, resulting from b), estimating a partial pressure of gas of interest dissolved in the blood of the user.

30. The method of claim 29, wherein c) comprises: (i) based on the partial pressure of gas of interest in the collecting chamber, resulting from b), estimating a partial pressure of transcutaneous gas of interest having diffused through the skin of the user; and (ii) based on the partial pressure of transcutaneous gas of interest resulting from (i), estimating a partial pressure of gas of interest dissolved in the blood of the user.

31. The method of claim 21, also comprising measuring a concentration of gas of interest in the ambient air, the concentration thus measured being taken into account in c).

32. The method of claim 21, comprising measuring an ambient air pressure, the ambient air pressure being taken into account in c).

33. The method of claim 21, wherein a) is implemented using a gas sensor, placed in the measuring chamber, and formed by an infrared radiation source and by a photodetector comprising two measurement channels configured to measure infrared radiation, emitted by the infrared radiation source, and having propagated through the measuring chamber in, respectively: a measurement spectral band that corresponds to an absorption spectral band of the gas of interest; and a reference spectral band, considered as not absorbed by the gas of interest; and wherein a) comprises a comparison of signals respectively generated by the two measurement channels; and solving a second-degree polynomial, the polynomial being defined by predetermined parameters as well as by the comparison of signals respectively generated by the two measurements channels, so as to estimate the concentration of gas of interest in the measuring chamber.

34. The method of claim 33, wherein the root of the polynomial is a non-integer power of the concentration of gas of interest in the measuring chamber.

35. The method of claim 33, wherein the parameters of the second-degree polynomial are defined in a supervised calibrating phase, using calibration samples having a known concentration of gas of interest.

36. The method of claim 33, wherein a) comprises measuring a water-vapor content in the measuring chamber, so that the water-vapor content is taken into account in the estimation of the concentration of gas of interest in the measuring chamber.

37. The method of claim 29, further comprising measuring the temperature in the measuring chamber and in the collecting chamber, wherein b) comprises taking into account the temperatures thus measured.

38. The method of claim 29, further comprising measuring the pressure in the measuring chamber and in the collecting chamber, wherein b) comprises taking into account the pressures thus measured.

39. The method of claim 29, wherein the gas of interest is carbon dioxide.

40. The method of claim 29, wherein the measurement spectral band is variable, so as to address various gases of interest.

Description

FIGURES

[0065] FIGS. 1A and 1B are overviews of the device.

[0066] FIGS. 2A and 2B show the main elements of the device.

[0067] FIGS. 3A and 3B schematically show the location of the measuring chamber and the measuring chamber, respectively.

[0068] FIG. 3C shows a detail of the far end of the device.

[0069] FIG. 4 illustrates the gas flows through the device.

[0070] FIG. 5 shows the main steps of a method implementing the device.

[0071] FIGS. 6A and 6B show a measuring chamber and an auxiliary measuring chamber, respectively.

[0072] FIGS. 7A and 7B show obtained experimental results.

DESCRIPTION OF PARTICULAR EMBODIMENTS

[0073] FIGS. 1A and 1B are general views of one example of a device 1 according to the invention. The device 1 is intended to be placed in contact with the skin of a human or animal user, or at a small distance from its skin, and typically at a distance smaller than 1 cm. It comprises a main body 2 and a fastening element 3, the latter being, in this example, a strap. The main body 2 comprises a contact face 4, intended to be placed on the skin. The contact face is substantially planar, in the sense that it lies parallel to an XY plane, certain segments possibly being inclined with respect to the XY plane. The main body also comprises a far end 6, opposite the contact face 4, and substantially parallel to the latter. The contact surface 4 and the far end 6 are connected to each other by a lateral face 5 that lies about a transverse axis Z that is perpendicular to the XY plane.

[0074] As shown in FIG. 1B, the contact face 4 bounds a collecting chamber 10. The latter is, in this example, bounded by a plate that forms part of the contact face 4, and in which through-apertures 11 are formed. The device also comprises a heating element 12 that allows the contact face 4 that bounds the collecting chamber 10 to be brought to a temperature above 37° C., and preferably comprised between 40° C. and 50° C., and preferably between 40° C. and 45° C., 42° C. for example. The heating element 12 is for example a resistor that is arranged on the plate bounding the collecting chamber 10, and that generates heat via Joule heating. Specifically, a local and moderate increase in temperature, in the vicinity of the skin, promotes a dilation of the blood capillaries, this increasing the diffusion of a transcutaneous gas of interest, through the skin. By gas of interest, what is meant is a gas a concentration of which in a living animal or human body, and more particularly in the blood, is to be determined. In the example described below, non-limitingly, the gas of interest is carbon dioxide, a content of which in the blood of the user it is sought to estimate. Inside the main body 2, FIG. 2A shows a duct 14, intended to guide the transcutaneous CO.sub.2 collected by the collecting chamber 10 to a measuring chamber 20, then downstream of the latter, to outlets 14.sub.s, as described below.

[0075] In order to facilitate the transport of transcutaneous CO.sub.2 to the measuring chamber 20, the main body 2 comprises apertures 15, intended to admit ambient air. Ambient air designates the air extending around the device 1. In the example shown in FIG. 2B, the apertures 15 are produced through the lateral face 5. Under the effect of the heat given off in the collecting chamber 10 under the action of the heating element 12, a thermal convection current is formed. Therefore, the transcutaneous CO.sub.2 and ambient air admitted into the main body form a gas mixture that moves away from the contact face 4 and propagates toward the measuring chamber 20, in the direction of the end 6. Thus, the ambient air admitted into the main body 2 through the apertures 15 acts as a carrier gas and transports the transcutaneous CO.sub.2 through the main body, and notably through the collecting chamber 10, to the measuring chamber 20. The lateral apertures, and the heating applied in the collecting chamber, form a chimney effect, which acts as a “motor” allowing a propagation of CO.sub.2 through the main body 2 of the device, via convection.

[0076] The measuring chamber 20 is, in this example, placed substantially parallel to the contact face 4. The function of the measuring chamber is to estimate a CO.sub.2 concentration of the gas mixture flowing through the main body 2. To this end, the measuring chamber 20 comprises a gas sensor 23. A plurality of types of sensors may be used to this end, for example optical sensors or electrochemical sensors, the latter notably being able to be based on metal oxides (MOX sensors). The inventors have estimated that it is preferable to use an optical sensor, and more precisely an infrared sensor. Such a sensor does not require any particular maintenance, and is particularly compact, and inexpensive. In addition, such a sensor is very specific with respect to characterization of chemical bonds. It is suitable for detecting small molecules, carbon dioxide for example. Gas sensors based on absorption of infrared light are known, and are usually designated by the term NDIR sensors (NDIR being the acronym of non-dispersive infrared). This type of sensor comprises an infrared radiation source 21 that generally emits in a spectral band comprised between 1 μm and 20 μm. It also comprises a photodetector 22 that is sensitive to the infrared radiation. The principle is based on the attenuation, by the analyzed gas, of the infrared radiation emitted by the source. The infrared source 21 and the photodetector 22 form the gas sensor 23, the operating principle of which is described in detail with reference to FIGS. 6A and 6B. The photodetector 22 is for example a thermopile.

[0077] The device preferably comprises an auxiliary measuring chamber 30. The auxiliary measuring chamber 30 comprises an auxiliary CO.sub.2 sensor 33 that operates in the same way as the sensor 23 placed in the measuring chamber 20. It comprises an infrared radiation source 31 and a photodetector 32. The auxiliary sensor 33 is intended to estimate a concentration of CO.sub.2 in the ambient air surrounding the main body 2. The use of such a sensor is preferable, but not essential.

[0078] The device comprises circuit boards 18.sub.1, 18.sub.2, 18.sub.3. All of the circuit boards form a processing and computing unit 18, configured to drive the various electronic components of the device, and to process the data output by the various detectors. The processing and computing unit may comprise computing means, a microprocessor for example, integrated into the device or located remotely from the latter and connected thereto by a wired or wireless link. The processing and computing unit is also configured to implement the method for estimating the CO.sub.2 content dissolved in the blood of the user, the main steps of which are schematically shown in FIG. 5. For this purpose, the processing unit may comprise one or more microprocessors.

[0079] FIG. 3A shows the duct 14, which allows the gas mixture to be directed to the measuring chamber 20. Downstream of the measuring chamber 20, the gas mixture is delivered to exhaust apertures 16, which are shown in FIG. 3C. More precisely, the duct 14 is divided into two elementary ducts 14.sub.1 and 14.sub.2, which open onto the exhaust apertures 16 placed downstream of the outlets 14.sub.s. FIG. 3B shows the measuring chamber 20, which especially comprises the photodetector 22.

[0080] At the far end 6, apertures 36 are produced that allow ambient air to gain access to an auxiliary measuring chamber 30. The auxiliary measuring chamber 30 is isolated from the rest of the components of the main body, such that it is exposed only to the ambient air extending around the main body 2. The auxiliary measuring chamber does not receive transcutaneous CO.sub.2. The auxiliary measuring chamber 30 is optional. It allows the precision with which the partial pressure of transcutaneous CO.sub.2 in the gas mixture flowing through the measuring chamber 20 may be estimated to be refined.

[0081] FIG. 4 schematically shows the fluidic currents formed in the device. Transcutaneous CO.sub.2 is admitted into the device through the through-apertures 11 formed in the contact face 4, which through-apertures open into the collecting chamber 10 (see arrow F.sub.1). Ambient air is admitted through the apertures 15 formed in the side wall 5 (see arrows F.sub.2). The gas mixture comprising the ambient air and transcutaneous CO.sub.2 propagates to the measuring chamber 20 (see arrow F.sub.3), through the duct 14. The gas mixture then extends into the measuring chamber 20, between the infrared radiation source 21 and the photodetector 22 (see arrow F.sub.4), before leaving the main body 2 (see arrows F.sub.5 and F.sub.6).

[0082] In parallel to this flow, ambient air diffuses into the auxiliary measuring chamber 30 through the apertures 36 (arrows F.sub.10), and extends, by diffusion, through the auxiliary measuring chamber (arrows F.sub.11 and arrows F.sub.12).

[0083] The processing and computing unit 18 collects the measurements of the partial pressure of CO.sub.2 in the measuring chamber 20 and in the auxiliary measuring chamber 30. On the basis of these measurements, the partial pressure of transcutaneous CO.sub.2, and the partial pressure of CO.sub.2 in the blood, are successively estimated, via the steps schematically shown in FIG. 5.

[0084] Step 100: determining the concentration of CO.sub.2 in the measuring chamber 20. Measurement of the partial pressure of CO.sub.2 in the measuring chamber 20 is a particularly important step. FIG. 6A schematically shows the measuring chamber 20. It defines a measurement volume that extends between an infrared radiation source 21 and the photodetector 22. The source and detector are components of the CO.sub.2 sensor 23.

[0085] The photodetector 22 comprises a measurement channel 22.sub.1 and a reference channel 22.sub.2. The measurement channel 22.sub.1 is configured to detect radiation having propagated through the measuring chamber 20, in an absorption spectral band that corresponds to the gas species that it is desired to analyze, in the present case CO.sub.2. Thus, the measurement channel 22.sub.1 performs a measurement of the intensity of the radiation transmitted by the gas mixture in a detection spectral band centered on λ.sub.1=4.26 μm. The reference channel 22.sub.2 is configured to detect radiation having propagated through the measuring chamber 20 in a reference spectral band, in which the absorption by the gas mixture is considered to be negligible. The reference spectral band is for example centered on λ.sub.1=3.91 μm. It will be noted that at the detection wavelength λ.sub.1 and at the reference wavelength λ.sub.2, the absorption of the radiation by water vapor may be considered to be equal. Otherwise, moisture concentration is taken into account in the model, as described below. The same goes for the absorption of air, which may be considered to be equal at the two wavelengths.

[0086] The measuring chamber 20 comprises, preferably, a temperature sensor 24, a humidity sensor 25, and a pressure sensor 26. The auxiliary measuring chamber 30 comprises, preferably, an auxiliary temperature sensor 34, an auxiliary humidity sensor 35, and an auxiliary pressure sensor 36.

[0087] Let U.sub.0(λ.sub.1) and U(λ.sub.1) be the voltages measured across the terminals of the measurement channel 22.sub.1, in the absence of gas from and in the presence of gas in the measuring chamber 20, respectively. Let U.sub.0(λ.sub.2) and U(λ.sub.2) be the voltages measured across the terminals of the measurement channel 22.sub.2, in the absence of gas from and in the presence of gas in the measuring chamber 20, respectively. If C.sub.CO2 is the molar concentration of CO.sub.2 in the gas mixture, it may be shown that:

[00001]$\begin{array}{cc}-\ln \left[\frac{\frac{U\left({\lambda }_1\right)}{U\left({\lambda }_2\right)}}{\frac{U_0\left({\lambda }_1\right)}{U_0\left({\lambda }_2\right)}}\right]=\left[k_{\mathrm{CO}2}\left({\lambda }_1\right)-k_{\mathrm{CO}2}\left({\lambda }_2\right)\right]C_{\mathrm{CO}2}+\left[k_{H2O}\left({\lambda }_1\right)-k_{H2O}\left({\lambda }_2\right)\right]C_{H2O}+\left[A_{\mathrm{air}}\left({\lambda }_1\right)-A_{\mathrm{air}}\left({\lambda }_2\right)\right]& \left(1\right)\end{array}$

where:

[0088] k.sub.CO2 is the attenuation coefficient of CO.sub.2;

[0089] k.sub.H2O is the attenuation coefficient of water, i.e. H.sub.2O, vapor;

[0090] A.sub.air is the attenuation of air;

[0091] C.sub.H2O is the molar concentration of H.sub.2O (water vapor) in the gas mixture.

[0092] However, it may be considered that A.sub.air(λ.sub.1)=A.sub.air(λ.sub.2) et k.sub.H2O(λ.sub.1)=k.sub.H2O (λ.sub.2) and that k.sub.CO2(λ.sub.2)=0.

Expression (1) then becomes:

[00002]$\begin{array}{cc}-\ln \left[\frac{\frac{U\left({\lambda }_1\right)}{U\left({\lambda }_2\right)}}{\frac{U_0\left({\lambda }_1\right)}{U_0\left({\lambda }_2\right)}}\right]\approx k_{\mathrm{CO}2}\left({\lambda }_1\right)C_{\mathrm{CO}2}& \left(2\right)\end{array}$

[0093] Thus, to a first approximation, the concentration of CO.sub.2 in the measuring chamber 20 may be estimated via a comparison, taking the form of a ratio, between the signals U(λ.sub.1) and U(λ.sub.2) generated by the measurement photodetector and the reference photodetector, respectively, the ratio

[00003]$\frac{U_0\left({\lambda }_1\right)}{U_0\left({\lambda }_2\right)}$

having been determined by calibration prior to use of the device, and being considered to remain constant during use of the device. Differential measurement is then spoken of, because the CO.sub.2 concentration is estimated via a simple comparison of the signals U(λ.sub.1) and U(λ.sub.2). Expression (2) corresponds to the application of a linear model, the attenuation being considered to vary linearly as a function of CO.sub.2 concentration.

[0094] However, such a model does not take into account a number of causes of uncertainty or variability, and in particular the fact that:

[0095] the concentration C.sub.H2O of water vapor in the gas mixture is much higher than the concentration of CO.sub.2;

[0096] because of multiple reflections inside the measuring chamber, the distance traveled by the detected infrared radiation, between the infrared radiation source 21 and the photodetector 22, is subject to variability. Specifically, the internal walls of the measuring chamber are preferably reflective. Consequently, the distances traveled by the light rays emitted by the source through the enclosure are variable. Because of the variability in the optical paths traveled through the gas, the attenuation lies in a range of variation. FIG. 6A shows various optical paths, schematically represented by dashed arrows. The linear model such as described with reference to expression (2) does not take into account such a variability in the optical paths.

[0097] the detection spectral band is not perfectly monochromatic, but extends over a certain spectral range Δλ.sub.1 about the detection wavelength λ.sub.1. However, in the addressed spectral range, the attenuation coefficients of CO.sub.2, k.sub.CO2 and of water vapor, k.sub.H2O vary.

[0098] These imperfections form sources of non-linearity. For this reason the inventors consider that the linear model, such as described by expressions (1) and (2), may be improved. They propose to model the measurements taken by the photodetectors with a quadratic linear model dependent on a non-integer power of the concentration of the gases. According to such a model, the measurements carried out by the detection and reference photodetectors are modeled by the following expression:

[00004]$\begin{array}{cc}-\ln \left[\frac{\frac{U\left({\lambda }_1\right)}{U\left({\lambda }_2\right)}}{\frac{U_0\left({\lambda }_1\right)}{U_0\left({\lambda }_2\right)}}\right]=m_1{C}_{\mathrm{CO}2}^{u_1}+n_1{C}_{\mathrm{CO}2}^{2u_1}+n_2{C}_{H2O}^{u_2}+n_2{C}_{H2O}^{2u_2}+{\mathrm{qC}}_{\mathrm{CO}2}^{u_1}.Math.C_{H2O}^{u_2}& \left(3\right)\end{array}$

where:

[0099] C.sub.CO2, and C.sub.H2O are the concentrations of CO.sub.2 and H.sub.2O in the gas mixture, in the measuring chamber 20; C.sub.H2O may be obtained by implementing a humidity sensor 25 in the measuring chamber 20;

[0100] m.sub.1, n.sub.1, m.sub.2, n.sub.2, and q are scalar parameters of the model;

[0101] u.sub.1, u.sub.2 are real numbers, which may be non-integers, corresponding to the powers of the concentrations;

[0102] the ratio U.sub.0(λ.sub.1)/U.sub.0(λ.sub.2) is determined by calibration prior to use of the device.

The bilinear term C.sub.CO2.sup.u.sup.1.Math.C.sub.H2O.sup.u.sup.2 may be neglected, in which case q=0.

[0103] Taking into account such a model has the advantage that it may be inverted analytically in a simple way, without requiring powerful computing means.

[0104] The concentration C.sub.H2O is determined by way of the humidity sensor 25. It is therefore a quantity that is considered to be known.

[0105] By making the following changes of variables:

[00005]$\begin{array}{cc}\gamma =m_2{C}_{H2O}^{u_2}+n_2{C}_{H2O}^{2u_2}+\ln \left[\frac{\frac{U\left({\lambda }_1\right)}{U\left({\lambda }_2\right)}}{\frac{U_0\left({\lambda }_1\right)}{U_0\left({\lambda }_2\right)}}\right]& \left(4\right)\\ \beta ={\mathrm{qC}}_{H2O}^{u_2}+m_1& \left(5\right)\\ x=C_{\mathrm{CO}2}^{u_1}& \left(6\right)\end{array}$

[0106] Expression (3) may be written:

n.sub.1x.sup.2+βx+γ=0  (7)

[0107] It is a polynomial of the second degree, the solutions of which are:

[00006]$\begin{array}{cc}x=\frac{-\beta \pm \sqrt{{\beta }^2-4n_1\gamma }}{2n_1}& \left(8\right)\\ \mathrm{and}{C}_{\mathrm{CO}2}=\sqrt[u_1]{x}& \left(9\right)\end{array}$

[0108] Equation (8) gives two possible values of x, only one of which, generally the lowest, allows a C.sub.CO2 concentration comprised in a predetermined range of values to be obtained. The predetermined range of values corresponds to values that are to be expected given the targeted application.

[0109] Use of the humidity sensor 25 is preferable, but not essential. According to one simplified model, the measurements may be modeled by:

[00007]$\begin{array}{cc}-\ln \left[\frac{\frac{U\left({\lambda }_1\right)}{U\left({\lambda }_2\right)}}{\frac{U_0\left({\lambda }_1\right)}{U_0\left({\lambda }_2\right)}}\right]=m_1{C}_{\mathrm{CO}2}^{u_1}+n_1{C}_{\mathrm{CO}2}^{2u_1}& \left(3^{\prime }\right)\end{array}$

[0110] By making the changes of variables:

[00008]$\begin{array}{cc}\gamma =\ln \left[\frac{\frac{U\left({\lambda }_1\right)}{U\left({\lambda }_2\right)}}{\frac{U_0\left({\lambda }_1\right)}{U_0\left({\lambda }_2\right)}}\right]& \left(4^{\prime }\right)\\ \beta =m_1& \left(5^{\prime }\right)\\ x=C_{\mathrm{CO}2}^{u_1}& \left(6^{\prime }\right)\end{array}$

[0111] Expression (3) may be written:

n.sub.1x.sup.2+βx+γ=0  (7′)

[0112] It is a polynomial of the second degree, the solutions of which are:

[00009]$\begin{array}{cc}x=\frac{-\beta \pm \sqrt{{\beta }^2-4n_1\gamma }}{2n_1}& \left(8^{\prime }\right)\end{array}$

[0113] It will be noted that such a model may be used to measure CO.sub.2, or any other gas species an absorption spectral band of which is known, independently of the device and application described in this patent application. Thus, such a model may be used to process measurements based on the attenuation of infrared radiation by a gas species, the measurements being taken both in the absorption spectral band of the gas species in question, and in a reference spectral band in which the absorption of the radiation is considered to be negligible.

[0114] Step 110 Determining the partial pressure of CO.sub.2 in the measuring chamber 20.

[0115] Having determined the molar concentration C.sub.CO2, the partial pressure of CO.sub.2 in the measuring chamber 20 is determined by applying the ideal gas law:

P.sub.CO2.sup.20=C.sub.CO2RT.sup.20  (10)

where:

[0116] T.sup.20 corresponds to the temperature in the measuring chamber 20, measured using the temperature sensor 24;

[0117] P.sub.CO2.sup.20 is the partial pressure of CO.sub.2 in the measuring chamber 20;

[0118] R is the ideal gas constant.

[0119] Step 120. Determining the partial pressure of CO.sub.2 in the collecting chamber 10. On the basis of the partial pressure of CO.sub.2 in the measuring chamber 20, the partial pressure of CO.sub.2 in the collecting chamber 10 is determined. To do this, the following expression is used, this expression relating the total pressure measured in the measuring chamber 20, which is denoted P.sub.Total.sup.20, to the total pressure measured in the collecting chamber 10, which is denoted P.sub.Total.sup.10.

[00010]$\begin{array}{cc}{P}_{\mathrm{CO}2}^{10}=\frac{P_{\mathrm{Total}}^{\mathrm{col}}}{P_{\mathrm{Total}}^{20}}{P}_{\mathrm{CO}2}^{20}& \left(11\right)\end{array}$

where:

[0120] P.sub.Total.sup.20 is the total pressure measured in the measuring chamber 20 by the pressure sensor 26.

[0121] P.sub.Total.sup.col is the total pressure measured in the collecting chamber 10 by a pressure sensor 19 placed in the collecting chamber 10, as shown in FIG. 4.

[0122] Step 130 Determining Transcutaneous CO.sub.2 Pressure.

[0123] This step aims to determine the transcutaneous CO.sub.2 pressure on the basis of the partial pressure of CO.sub.2 in the collecting chamber 10. The arrival of air through the lateral apertures 15 must be taken into account. The collecting chamber 10 comprises a temperature sensor 17. The device also comprise an ambient-temperature sensor 34. When the device comprises an auxiliary measuring chamber 30, the ambient-temperature sensor is preferably placed in the latter.

[0124] According to a first approach, the CO.sub.2 in the ambient air admitted through the lateral apertures 15 is neglected. It is thus assumed that the CO.sub.2 flowing through the device is very predominantly transcutaneous CO.sub.2.

[0125] According to this first approach, the transcutaneous CO.sub.2 pressure P.sub.CO.sub.2.sup.tc in the collecting chamber 10 is:

[00011]$\begin{array}{cc}{P}_{{\mathrm{CO}}_2}^{\mathrm{tc}}\left(T^{\mathrm{tc}}\right)=\frac{T^{\mathrm{tc}}}{T^{10}}.Math.g_{{\mathrm{CO}}_2}^{\mathrm{tc}/\mathrm{col}}.Math.P_{{\mathrm{CO}}_2}^{\mathrm{col}}\left(T^{10}\right)& \left(12\right)\end{array}$

where:

[0126] P.sub.CO2.sup.tc (T.sup.tc) is the transcutaneous CO.sub.2 pressure at the transcutaneous temperature T.sup.tc, i.e. the skin temperature, and expressed in mmHg. This temperature may be measured by a temperature sensor 17′, as described below.

[0127] P.sub.CO2.sup.col(T.sup.10) is the CO.sub.2 pressure measured at the temperature T.sup.10 in the collecting chamber and expressed in mmHg;

[0128] T.sup.10 is a temperature measured in the collecting chamber by a temperature sensor 17.

[0129] The notation X (T.sup.x) denotes the fact that a quantity X is measured at the temperature T.sup.x.

[0130] The skin temperature T.sup.tc may be assumed to be known or be measured by a skin-temperature sensor 17′, allowing the temperature of the skin to be known. The skin-temperature sensor may also allow heating of the skin with respect to a setpoint temperature, the latter being 42° C., to be regulated.

g.sub.CO.sub.2.sup.tc/col is a correction factor allowing a dilution, in the collecting chamber 10, induced by the air admitted through the apertures 15 to be taken into account. This factor may be determined via the following expression:

[00012]$\begin{array}{cc}{g}_{{\mathrm{CO}}_2}^{\mathrm{tc}/\mathrm{col}}=\frac{k_{\mathrm{blood}}.Math.{\rho }_{\mathrm{skin}}^{\mathrm{blood}}.Math.S_{\mathrm{skin}}.Math.H_{b:\mathrm{air}}+D_{\mathrm{air}}}{k_{\mathrm{blood}}.Math.{\rho }_{\mathrm{skin}}^{\mathrm{blood}}.Math.S_{\mathrm{skin}}.Math.H_{b:\mathrm{air}}}& \left(13\right)\end{array}$

[0131] k.sub.blood is a coefficient of mass transfer, expressed in m.Math.s.sup.−1, between blood and air;

[0132] S.sub.blood is the area of exchange at the blood/air interface

[0133] ρ.sub.skin.sup.blood is a ratio of blood perfusion to the surface of the skin:

[00013]$\begin{array}{cc}{\rho }_{\mathrm{skin}}^{\mathrm{blood}}=\frac{S_{\mathrm{blood}}}{S_{\mathrm{skin}}}& \left(14\right)\end{array}$

[0134] S.sub.blood is the area of blood involved in the emission of transcutaneous CO.sub.2;

[0135] S.sub.skin is the area of the skin involved in the emission of transcutaneous CO.sub.2: it is a question of the area of skin placed facing the device;

[0136] H.sub.b:air is a blood/air partition coefficient also corresponding to a Henry's law constant between a liquid concentration and a gas concentration;

[0137] D.sub.air is the volumetric flow rate of air through the device (m.sup.3/s), through the lateral apertures 15—it is established as described below:

[0138] The coefficient of blood/air mass transfer and the blood/air partition coefficient are physiological quantities given in the literature: For example, the coefficient of mass transfer k.sub.blood may be considered to be equal to 23.1 μm/s. The blood/air partition coefficient H.sub.b:air may be, for example, considered to be equal to 0.59 at the temperature of 37° C. The ratio ρ.sub.skin.sup.blood is adjusted depending on the bodily region to which the device 1 is applied.

[0139] The area S.sub.skin of the skin placed facing the device is defined by the geometry of the device.

[0140] This equation may be written:

[00014]$\begin{array}{cc}{g}_{{\mathrm{CO}}_2}^{\mathrm{tc}/\mathrm{col}}=1+\frac{D_{\mathrm{air}}}{D_{{\mathrm{CO}}_2}^{b:\mathrm{air}}}& \left(16\right)\end{array}$

where
D.sub.CO.sub.2.sup.b:air is the rate of diffusion of CO.sub.2 through the blood/air interface. It may be established using the following expression:

D.sub.CO2.sup.b:air=ρ.sub.skin.sup.blood.Math.S.sub.skin.Math.k.sub.blood.Math.H.sub.b:air  (17)

and

[00015]$\begin{array}{cc}{D}_{\mathrm{air}}\left(T^{10}\right)=R.Math.T^{10}\frac{\left(P_{\mathrm{air}}^{\mathrm{out}\mathrm{col}}\left(T^{10}\right)-P_{\mathrm{air}}^{\mathrm{amb}}\left(T^{\mathrm{amb}}\right)\right)}{P_{\mathrm{air}}^{\mathrm{out}\mathrm{col}}\left(T^{10}\right).Math.R^{\mathrm{col}-\mathrm{amb}}}& \left(18\right)\end{array}$

[0141] p.sub.air.sup.out col is the air pressure measured in the collecting chamber 10;

[0142] p.sub.air.sup.amb is the ambient air pressure, measured for example in the auxiliary chamber 30;

[0143] R.sup.col-amb is the Hagen-Poiseuille flow resistance between the collecting chamber and the ambient air;

[0144] R is the ideal gas constant.

The Hagen-Poiseuille flow resistance R.sup.col-amb may be estimated by making air flow at a set flow rate through the device on a testbed, and by measuring the pressures P.sub.air.sup.out col(T.sup.10) and P.sub.air.sup.amb(T.sup.amb) using the pressure sensors of the device. Then:

[00016]$\begin{array}{cc}{R}^{\mathrm{col}-\mathrm{amb}}=R.Math.T^{10}\frac{\left(P_{\mathrm{air}}^{\mathrm{out}\mathrm{col}}\left(T^{10}\right)-P_{\mathrm{air}}^{\mathrm{amb}}\left(T^{\mathrm{amb}}\right)\right)}{P_{\mathrm{air}}^{\mathrm{out}\mathrm{col}}\left(T^{10}\right).Math.D_{\mathrm{air}}\left(T^{10}\right)}& \left(19\right)\end{array}$

[0145] It may be shown that:

[CO.sub.2.sup.tc](T.sup.tc)≅g.sub.CO.sub.2.sup.tc/col.Math.[CO.sub.2.sup.10](T.sup.10)=g.sub.CO.sub.2.sup.tc/col.Math.[CO.sub.2.sup.20](T.sup.20)  (20)

[0146] [CO.sub.2.sup.tc] is the relative concentration of transcutaneous CO.sub.2 in the collecting chamber 10, expressed in ppm;

[0147] [CO.sub.2.sup.10] is the total concentration of CO.sub.2 in the collecting chamber 10, expressed in ppm;

[0148] [CO.sub.2.sup.20] is the concentration of CO.sub.2 in the measuring chamber 20;

[0149] [CO.sub.2.sup.amb] is the concentration of CO.sub.2 in the ambient air, measured in the auxiliary chamber 30, and expressed in ppm.

[0150] According to a second approach, geometric effects and temperature variations in the ambient air are taken into account. The collecting chamber 10 comprises a temperature sensor 17. The device also comprises an ambient-temperature sensor 34. When the device comprises an auxiliary measuring chamber 30, the ambient-temperature sensor is preferably placed in the latter. According to this second approach,

[00017]$\begin{array}{cc}{P}_{{\mathrm{CO}}_2}^{\mathrm{tc}}\left(T^{\mathrm{tc}}\right)=g_{{\mathrm{CO}}_2}^{\mathrm{tc}/\mathrm{col}}.Math.\frac{P_{{\mathrm{CO}}_2}^{\mathrm{out}\mathrm{col}}\left(T^{10}\right)}{P_{\mathrm{air}}^{\mathrm{out}\mathrm{col}}\left(T^{10}\right)}.Math.P_{\mathrm{air}}^{\mathrm{amb}}\left(T^{\mathrm{amb}}\right).Math.\frac{T^{\mathrm{tc}}}{T^{\mathrm{amb}}}.Math.\sqrt{\frac{S^{15}}{S^{11}}}& \left(21\right)\end{array}$

[0151] T.sup.amb is the temperature of the ambient air. It is measured by a temperature sensor 34, the latter being able to be placed in the auxiliary measuring chamber 30 when the device is equipped with such a chamber;

[0152] T.sup.tc is the skin temperature, for example equal to 42° C.;

[0153] T.sup.10 is the temperature in the collecting chamber, through which the transcutaneous CO.sub.2 passes; it is measured by the temperature sensor 17.

[0154] P.sub.air.sup.amb is the pressure of the ambient air. It is measured by a sensor 36 of ambient air pressure, said sensor being able to be placed in the auxiliary measuring chamber 30 when the device is equipped with such a chamber;

[0155] S.sup.15 is the area of the lateral apertures 15, through which the ambient air is admitted into the collecting chamber 10;

[0156] S.sup.11 is the area of the through-apertures 11, through which the transcutaneous CO.sub.2 is admitted into the collecting chamber 10.

[0157] According to a third approach, the CO.sub.2 concentration transported by the ambient air admitted into the collecting chamber 10 through the lateral apertures 15 is also taken into account. The CO.sub.2 concentration may be predetermined, i.e. reflect a predefined concentration, for example an average CO.sub.2 concentration in air. Advantageously, the concentration of CO.sub.2 in the ambient air is measured using the auxiliary measuring chamber 30, which was mentioned above, or via another measuring method.

[0158] The auxiliary measuring chamber contains only ambient air. FIG. 6B schematically shows the auxiliary measuring chamber. The operation of the auxiliary measuring chamber 30 may be similar to the operation of the measuring chamber 20. In particular, the auxiliary infrared radiation source 31 emits radiation in an emission spectral band comprised between 1 μm and 20 μm. The auxiliary photodetector 32 comprises a measurement channel 32.sub.1 centered on 4.28 μm and a reference channel 32.sub.2 centered on 3.91 μm. The auxiliary measuring chamber allows measurements to be obtained that may be modeled with a non-linear model such as described with reference to step 100:

[00018]$\begin{array}{cc}-\ln \left[\frac{\frac{U^{\prime }\left({\lambda }_1\right)}{U^{\prime }\left({\lambda }_2\right)}}{\frac{U_0^{\prime }\left({\lambda }_1\right)}{U_0^{\prime }\left({\lambda }_2\right)}}\right]=m_1^{\prime C_{\mathrm{CO}2}^{{\mathrm{amb}}^{u_1}}}+n_1^{\prime C_{\mathrm{CO}2}^{{\mathrm{amb}}^{2u_1}}}+m_2^{\prime C_{H2O}^{{\mathrm{amb}}^{u_2}}}+n_2^{\prime C_{H2O}^{{\mathrm{amb}}^{2u_2}}}+q^{\prime C_{\mathrm{CO}2}^{{\mathrm{amb}}^{u_1}}}.Math.C_{H2O}^{{\mathrm{amb}}^{u_2}}& \left(22\right)\end{array}$

where:

[0159] C.sub.CO2.sup.amb is the concentration of CO.sub.2 in the ambient air;

[0160] C.sub.H2O.sup.amb is the concentration of H.sub.2O in the ambient air, said concentration possibly being measured by an auxiliary humidity sensor 35 placed in the auxiliary measuring chamber 30;

[0161] U′.sub.0(λ.sub.1) and U′(λ.sub.1) are the signals, for example the voltages, measured across the terminals of the measurement channel 32.sub.1 in the absence of gas from and in the presence of gas in the auxiliary measuring chamber 30, respectively.

[0162] U′.sub.0(λ.sub.2) and U′(λ.sub.2) are the signals, for example the voltages, measured across the terminals of the reference channel 32.sub.2 in the absence of gas from and in the presence of gas in the auxiliary measuring chamber 30, respectively.

[0163] m′.sub.1′, n′.sub.1, m′.sub.2, n′.sub.2, and q′ are scalar parameters of the model corresponding to the auxiliary sensor 33;

[0164] u′.sub.1, u′.sub.2 are real numbers, which may be non-integers, corresponding to the powers of the concentrations.

[0165] The values m′.sub.1′, n′.sub.1, m′.sub.2, n′.sub.2, q′, u′.sub.1 and u′.sub.2 may be determined in a calibrating phase, such as that described in step 90.

[0166] It is considered that the partial pressure of CO.sub.2 in the blood is likely to be most accurately estimated when the concentration of CO.sub.2 in the ambient air is taken into account.

[0167] According to this third approach, the contribution due to CO.sub.2 in the ambient air having been admitted into the collecting chamber 10 through the lateral apertures is subtracted.

[00019]$\mathrm{Thus}:\begin{array}{cc}\left[{\mathrm{CO}}_2^{\mathrm{tc}}\right]\left(T^{\mathrm{tc}}\right)=g_{{\mathrm{CO}}_2}^{\mathrm{tc}/\mathrm{col}}.Math.\left[{\mathrm{CO}}_2^{10}\right]\left(T^{10}\right)-\frac{D_{\mathrm{air}}}{k_{\mathrm{blood}}.Math.{\rho }_{\mathrm{skin}}^{\mathrm{blood}}.Math.S_{\mathrm{skin}}.Math.H_{b:\mathrm{air}}}.Math.\sqrt{\frac{S^{15}}{S^{11}}}.Math.\left[{\mathrm{CO}}_2^{\mathrm{amb}}\right]\left(T^{\mathrm{amb}}\right)\mathrm{However},& \left(25\right)\\ \left[{\mathrm{CO}}_2^{10}\right]\left(T^{10}\right)=\left[{\mathrm{CO}}_2^{20}\right]\left(T^{20}\right)\mathrm{Hence}:& \left(26\right)\\ \left[{\mathrm{CO}}_2^{\mathrm{tc}}\right]\left(T^{\mathrm{tc}}\right)=g_{{\mathrm{CO}}_2}^{\mathrm{tc}/\mathrm{col}}.Math.\left[{\mathrm{CO}}_2^{20}\right]\left(T^{20}\right)-\frac{D_{\mathrm{air}}\left(T^{10}\right)}{D_{{\mathrm{CO}}_2}^{b:\mathrm{air}}}.Math.\sqrt{\frac{S^{15}}{S^{11}}}.Math.\left[{\mathrm{CO}}_2^{\mathrm{amb}}\right]\left(T^{\mathrm{amb}}\right)& \left(27\right)\end{array}$

D.sub.CO2.sup.b:air is the volumetric flow rate of diffusion of CO.sub.2 through the blood/air interface. If only the blood/air phase change is taken into account, we may use the expression:

D.sub.CO.sub.2.sup.b:air=ρ.sub.skin.sup.blood.Math.S.sub.skin.Math.k.sub.blood.Math.H.sub.b:air  (28)

[0168] If we also want to take into account the effect of the skin on the diffusion of the gas, we may introduce the binary diffusion coefficient of the skin D.sub.skin.sup.bin (m.Math.s.sup.−1). It defines the rate of diffusion of carbon dioxide through the skin.

In this case, the expression for D.sub.CO2.sup.b:air becomes:

D.sub.CO.sub.2.sup.b:air=ρ.sub.skin.sup.blood.Math.S.sub.skin.Math.K.sub.CO.sub.2.sup.blood to col  (29)

K.sub.CO.sub.2.sup.blood to col is the coefficient of total mass transfer between blood and air combining the phase change and the diffusion through the skin:

[00020]$\begin{array}{cc}\frac{1}{K_{{\mathrm{CO}}_2}^{\mathrm{blood}\mathrm{to}\mathrm{col}}}=\frac{1}{k_{\mathrm{blood}}{H}_{b:\mathrm{air}}}+\frac{1}{D_{\mathrm{skin}}^{\mathrm{bin}}}& \left(30\right)\end{array}$

D.sub.air(T.sup.10) is the volumetric flow rate of air admitted through the lateral apertures 15 into the collecting chamber 10, such that:

[00021]$\begin{array}{cc}{D}_{\mathrm{air}}\left(T^{10}\right)=R.Math.T^{10}\frac{\left(P_{\mathrm{air}}^{\mathrm{out}\mathrm{col}}\left(T^{10}\right)-P_{\mathrm{air}}^{\mathrm{amb}}\left(T^{\mathrm{amb}}\right)\right)}{P_{\mathrm{air}}^{\mathrm{out}\mathrm{col}}\left(T^{10}\right).Math.R^{\mathrm{col}-\mathrm{amb}}}& \left(31\right)\end{array}$

[0169] It is also possible to write:

[00022]$\begin{array}{cc}\left[{\mathrm{CO}}_2^{\mathrm{tc}}\right]\left(T^{\mathrm{tc}}\right)=g_{{\mathrm{CO}}_2}^{\mathrm{tc}/\mathrm{col}}.Math.\left[{\mathrm{CO}}_2^{10}\right]\left(T^{10}\right)-\frac{D_{\mathrm{air}}\left(T^{10}\right)}{D_{{\mathrm{CO}}_2}^{\mathrm{skin}}}.Math.\sqrt{\frac{S^{15}}{S^{11}}}.Math.\left[{\mathrm{CO}}_2^{\mathrm{amb}}\right]\left(T^{\mathrm{amb}}\right)\mathrm{However},& \left(32\right)\\ \left[{\mathrm{CO}}_2^{10}\right]\left(T^{10}\right)=\left[{\mathrm{CO}}_2^{20}\right]\left(T^{20}\right)\mathrm{Hence}& \left(33\right)\\ \left[{\mathrm{CO}}_2^{\mathrm{tc}}\right]\left(T^{\mathrm{tc}}\right)=g_{{\mathrm{CO}}_2}^{\mathrm{tc}/\mathrm{col}}.Math.\left[{\mathrm{CO}}_2^{20}\right]\left(T^{20}\right)-\frac{D_{\mathrm{air}}\left(T^{10}\right)}{D_{{\mathrm{CO}}_2}^{\mathrm{skin}}}.Math.\sqrt{\frac{S^{15}}{S^{11}}}.Math.\left[{\mathrm{CO}}_2^{\mathrm{amb}}\right]\left(T^{\mathrm{amb}}\right)& \left(34\right)\end{array}$

with:

[00023]$g_{{\mathrm{CO}}_2}^{\mathrm{tc}/\mathrm{col}}=1+\frac{D_{\mathrm{air}}\left(T^{10}\right)}{D_{{\mathrm{CO}}_2}^{b:\mathrm{air}}},$

as defined above.
D.sub.CO.sub.2.sup.skin is defined in relation to expression (48).

[0170] For the partial pressure, we have the relationship:

[00024]$\begin{array}{cc}{P}_{{\mathrm{CO}}_2}^{\mathrm{tc}}\left(T^{\mathrm{tc}}\right)=\frac{T^{\mathrm{tc}}}{T^{10}}\left[g_{{\mathrm{CO}}_2}^{\mathrm{tc}/\mathrm{col}}.Math.P_{{\mathrm{CO}}_2}^{10}\left(T^{10}\right)-\frac{D_{\mathrm{air}}{P}_{{\mathrm{CO}}_2}^{\mathrm{amb}}\left(T^{10}\right)}{k_{\mathrm{blood}}.Math.{\rho }_{\mathrm{skin}}^{\mathrm{blood}}.Math.S_{\mathrm{skin}}.Math.H_{b:\mathrm{air}}}\right]\mathrm{where}:& \left(35\right)\\ P_{{\mathrm{CO}}_2}^{\mathrm{tc}}\left(T^{\mathrm{tc}}\right)=\frac{T^{\mathrm{TC}}}{T^{10}}\left[P_{{\mathrm{CO}}_2}^{10}\left(T^{10}\right)+\frac{D_{\mathrm{air}}}{D_{{\mathrm{CO}}_2}^{b:\mathrm{air}}}.Math.\left[P_{{\mathrm{CO}}_2}^{10}\left(T^{10}\right)-P_{{\mathrm{CO}}_2}^{\mathrm{amb}}\left(T^{10}\right)\right]\right]& \left(36\right)\end{array}$

[0171] Step 140: Estimating the Partial Pressure of CO.sub.2 in the Blood.

[0172] This step aims to determine a function ƒ, such that:

P.sub.CO2.sup.blood=ƒ(P.sub.CO2.sup.tc)  (40)

where:

[0173] P.sub.CO2.sup.tc is the partial pressure of transcutaneous CO.sub.2 in step 130;

[0174] P.sub.CO2.sup.blood is the partial pressure of CO.sub.2 in the blood, this being the physiological quantity that the method is aimed at estimating.

[0175] In this step, it is assumed, according to a first variant, that the partial pressure of CO.sub.2 in the collecting chamber P.sub.CO2.sup.10 corresponds to the partial pressure of transcutaneous CO.sub.2 P.sub.CO2.sup.tc.

[0176] The partial pressure of CO.sub.2 in the blood may be estimated via an empirical formula given in J. W. Severinghaus “The current status of transcutaneous blood gas analysis and monitoring”, Blood Gas News, vol. 7, no 2, pp 4-9, 1998.

[00025]$\begin{array}{cc}{P}_{\mathrm{CO}2}^{\mathrm{blood}}\left(37°C.\right)=\frac{P_{\mathrm{CO}2}^{\mathrm{blood}}\left(T^{\mathrm{tc}}\right)-4}{e^{\left(0.046\left(T^{\mathrm{tc}}-37\right)\right)}}& \left(41\right)\end{array}$

where T.sup.tc is a temperature measured by the temperature sensor 17′.

[0177] When using formula (41), the temperatures are expressed in ° C. and the pressures are expressed in mmHg.

[0178] Other methods may be applied to estimate P.sub.CO2.sup.blood on the basis of P.sub.CO2.sup.tc. According to a first variant, this estimation may be based on modeling of the flow Φ.sub.CO2 of transcutaneous CO.sub.2 diffusing through the skin.

[0179] The gas flow Φ.sub.CO.sub.2.sup.g,blood to collection of CO.sub.2 through the skin is modeled by:

an equation of gas-phase diffusion through the skin (skin effect):

Φ.sub.CO.sub.2.sup.g,blood to collection=S.sub.blood.Math.D.sub.skin.sup.bin.Math.(C.sub.CO.sub.2.sup.g,blood−C.sub.CO.sub.2.sup.col)  (40)

D.sub.skin.sup.bin is a binary diffusion coefficient of the skin (m.Math.s.sup.−1) corresponding to the coefficient of mass transfer of carbon dioxide through the skin. It is determined either experimentally from skin biopsies, or numerically, the parameters of the model being trained on people for whom reference measurements of carbon-dioxide concentration have been taken using another technique for measuring carbon-dioxide concentration in the blood (blood samples, measurement of exhaled air, measurement of transcutaneous pressure using electrochemical sensors, etc.), the parameters being adjusted to make the pressures estimated by the device correspond to the measured reference pressures.
C.sub.CO.sub.2.sup.g,blood is a molar gas-phase concentration of CO.sub.2 in the blood (mol.Math.mm.sup.−3) in equilibrium with the liquid-phase concentration, according to Henry's law:

C.sub.CO.sub.2.sup.g,blood=σ.sub.CO.sub.2.Math.P.sub.CO.sub.2.sup.blood  (41)

P.sub.CO.sub.2.sup.blood is the partial pressure of CO.sub.2 in the blood, expressed in mmHg (millimeters of mercury), and σ.sub.CO.sub.2 is the Ostwald solubility coefficient, which may be such that:

[00026]$\begin{array}{cc}{\sigma }_{{\mathrm{CO}}_2}=3.3\times {10}^{-2}\frac{\mathrm{mM}}{\mathrm{mmHg}}& \left(42\right)\end{array}$

C.sub.CO.sub.2.sup.col,* is the molar concentration of CO.sub.2 that there would be in equilibrium in the collecting chamber if carbon dioxide were not removed by convection. It is also the concentration that exists in the collecting chamber against the skin. It is the transcutaneous concentration corresponding to the transcutaneous pressure.
C.sub.CO.sub.2.sup.col is the molar concentration of CO.sub.2 in the collecting chamber 10.

[0180] Because of the removal, by convection, of CO.sub.2 from the collecting chamber, the concentration C.sub.CO.sub.2.sup.col is different from the concentration C.sub.CO2.sup.col,*.

[0181] The equation of mass transfer between the liquid phase and the gas phase related solely to the phase change is written:

Φ.sub.CO.sub.2.sup.g,blood to collection=−S.sub.bloodk.sub.bloodH.sub.b:air.Math.(C.sub.CO.sub.2.sup.col−C.sub.CO.sub.2.sup.col,*)  (43)

[0182] By combining the two mass-transfer equations, the following is obtained:

Φ.sub.CO.sub.2.sup.blood to collection=−S.sub.bloodk.sub.CO.sub.2.sup.blood to col.Math.(C.sub.CO.sub.2.sup.col−C.sub.CO.sub.2.sup.g,blood)  (44)

where K.sub.CO.sub.2.sup.blood to col is the coefficient of total mass transfer combining the phase change and diffusion through the skin:

[00027]$\begin{array}{cc}\frac{1}{K_{{\mathrm{CO}}_2}^{\mathrm{blood}\mathrm{to}\mathrm{col}}}=\frac{1}{k_{\mathrm{blood}}{H}_{b:\mathrm{air}}}+\frac{1}{D_{\mathrm{skin}}^{\mathrm{bin}}}& \left(45\right)\end{array}$

[0183] The following expression may be deduced from the above:

[00028]$\begin{array}{cc}{P}_{{\mathrm{CO}}_2}^{\mathrm{blood}}\left(T^{\mathrm{tc}}\right)=\frac{T^{\mathrm{tc}}}{T^{10}}.Math.g_{{\mathrm{CO}}_2}^{\mathrm{blood}/\mathrm{col}}.Math.P_{{\mathrm{CO}}_2}^{10}\left(T^{10}\right)& \left(46\right)\end{array}$

g.sub.CO.sub.2.sup.blood/col is a correction factor that corrects for the effect of dilution in the collecting chamber, for the effect of diffusion through the skin, and for the effect of change of phase from blood to air.

[00029]$\begin{array}{cc}{g}_{{\mathrm{CO}}_2}^{\mathrm{blood}/\mathrm{col}}=\frac{{\rho }_{\mathrm{skin}}^{\mathrm{blood}}.Math.S_{11}.Math.K_{{\mathrm{CO}}_2}^{\mathrm{blood}\mathrm{to}\mathrm{col}}+D_{\mathrm{air}}}{{\rho }_{\mathrm{skin}}^{\mathrm{blood}}.Math.S_{11}.Math.K_{{\mathrm{CO}}_2}^{\mathrm{blood}\mathrm{to}\mathrm{col}}}& \left(47\right)\end{array}$

[0184] Consequently

[00030]$\begin{array}{cc}{P}_{{\mathrm{CO}}_2}^{\mathrm{blood}}\left(T^{\mathrm{tc}}\right)=P_{{\mathrm{CO}}_2}^{\mathrm{tc}}\left(T^{\mathrm{tc}}\right)+\frac{T^{\mathrm{tc}}}{T^{10}}.Math.\frac{D_{\mathrm{air}}}{D_{{\mathrm{CO}}_2}^{\mathrm{skin}}}{P}_{{\mathrm{CO}}_2}^{10}\left(T^{10}\right)& \left(48\right)\end{array}$

where:

[0185] D.sub.CO.sub.2.sup.skin is a volumetric flow rate of diffusion of CO.sub.2 through the skin, such that:

D.sub.CO.sub.2.sup.skin=ρ.sub.skin.sup.blood.Math.S.sub.skin.Math.D.sub.skin.sup.bin  (49)

[0186] D.sub.air(T.sup.10) is a volumetric flow rate of ambient air in the collecting chamber 10, such that:

[00031]$\begin{array}{cc}{D}_{\mathrm{air}}\left(T^{10}\right)=R.Math.T^{10}\frac{\left(P_{\mathrm{air}}^{\mathrm{out}\mathrm{col}}\left(T^{10}\right)-P_{\mathrm{air}}^{\mathrm{amb}}\left(T^{10}\right)\right)}{P_{\mathrm{air}}^{\mathrm{out}\mathrm{col}}\left(T^{10}\right).Math.R^{\mathrm{col}-\mathrm{amb}}}& \left(50\right)\end{array}$

[0187] P.sub.CO.sub.2.sup.amb(T.sup.10) is the CO.sub.2 pressure, in the collecting chamber, due to the ambient air, such that:

[00032]$\begin{array}{cc}{P}_{{\mathrm{CO}}_2}^{\mathrm{amb}}\left(T^{10}\right)=\frac{T^{10}}{T^{\mathrm{amb}}}.Math.P_{{\mathrm{CO}}_2}^{\mathrm{amb}}\left(T^{\mathrm{amb}}\right)& \left(51\right)\end{array}$

[0188] R.sup.col-amb corresponds to the fluid resistance between the ambient air and the collecting chamber.

[0189] The partial pressure P.sub.CO2.sup.blood estimated using the above formulas is estimated at the temperature of the collecting chamber 10, which is above 37° C. However, the blood is at a temperature of 37° C. It is therefore necessary to make a correction in order to obtain an estimate of P.sub.CO2.sup.blood. Such a correction may be made according to J. W. Severinghaus “The current status of transcutaneous blood gas analysis and monitoring”, Blood Gas News, vol. 7, no 2, pp 4-9, 1998, the latter defining a transfer function f, such that:

[00033]$\begin{array}{cc}f=\frac{P_{\mathrm{CO}2}^{\mathrm{blood}}\left(T^{10}\right)}{P_{\mathrm{CO}2}^{\mathrm{blood}}\left(T=37°C.\right)}=e^{0.046\left[T^{10}-37\right]}& \left(52\right)\end{array}$

[0190] where T.sup.10 is the temperature in the collecting chamber 10.

[0191] Thus:

[00034]$\begin{array}{cc}{P}_{{\mathrm{CO}}_2}^{\mathrm{blood}}\left(37°C.\right)=\frac{P_{{\mathrm{CO}}_2}^{\mathrm{blood}}\left(T^{10}\right)}{\exp \left(0.046\left[T^{10}-37\right]\right)}& \left(53\right)\end{array}$

[0192] The partial pressure output from the preceding equation is expressed in mmHg.

[0193] According to a second variant, the CO.sub.2 present in the ambient air admitted into the collecting chamber through the through-apertures is taken into account. The equations presented with reference to the first variant become:

[00035]$\begin{array}{cc}{P}_{{\mathrm{CO}}_2}^{\mathrm{tc}}\left(T^{\mathrm{tc}}\right)=\frac{T^{\mathrm{tc}}}{T^{10}}\left[g_{{\mathrm{CO}}_2}^{\mathrm{blood}/\mathrm{col}}.Math.P_{{\mathrm{CO}}_2}^{\mathrm{col}}\left(T^{10}\right)-\frac{D_{\mathrm{air}}.Math.P_{{\mathrm{CO}}_2}^{\mathrm{amb}}\left(T^{10}\right)}{K_{{\mathrm{CO}}_2}^{\mathrm{blood}\mathrm{to}\mathrm{col}}.Math.{\rho }_{\mathrm{skin}}^{\mathrm{blood}}.Math.S_{\mathrm{skin}}}\right]& \left(55\right)\end{array}$

g.sub.CO.sub.2.sup.blood/col is a correction factor that corrects for the effect of dilution in the collecting chamber, for the effect of diffusion through the skin, and for the effect of change of phase from blood to air.

[00036]$\begin{array}{cc}{g}_{{\mathrm{CO}}_2}^{\mathrm{blood}/\mathrm{col}}=\frac{{\rho }_{\mathrm{skin}}^{\mathrm{blood}}.Math.S_{\mathrm{skin}}.Math.K_{{\mathrm{CO}}_2}^{\mathrm{blood}\mathrm{to}\mathrm{col}}+D_{\mathrm{air}}}{{\rho }_{\mathrm{skin}}^{\mathrm{blood}}.Math.S_{\mathrm{skin}}.Math.K_{{\mathrm{CO}}_2}^{\mathrm{blood}\mathrm{to}\mathrm{col}}}& \left(56\right)\\ \mathrm{that}\mathrm{is}{g}_{{\mathrm{CO}}_2}^{\mathrm{blood}/\mathrm{col}}=1+\frac{D_{\mathrm{air}}}{{\rho }_{\mathrm{skin}}^{\mathrm{blood}}.Math.S_{\mathrm{skin}}.Math.K_{{\mathrm{CO}}_2}^{\mathrm{blood}\mathrm{to}\mathrm{col}}}& \left(57\right)\\ \mathrm{so}{g}_{{\mathrm{CO}}_2}^{\mathrm{blood}/\mathrm{col}}=1+\frac{D_{\mathrm{air}}}{{\rho }_{\mathrm{skin}}^{\mathrm{blood}}.Math.S_{\mathrm{skin}}}\left[\frac{1}{k_{\mathrm{blood}}{H}_{b:\mathrm{air}}}+\frac{1}{D_{\mathrm{skin}}^{\mathrm{bin}}}\right]& \left(58\right)\end{array}$

[0194] It may be deduced therefrom that:

[00037]$\begin{array}{cc}{P}_{{\mathrm{CO}}_2}^{\mathrm{blood}}\left(T^{\mathrm{tc}}\right)=P_{{\mathrm{CO}}_2}^{\mathrm{tc}}\left(T^{\mathrm{tc}}\right)+\frac{T^{\mathrm{tc}}}{T^{10}}.Math.\frac{D_{\mathrm{air}}}{D_{{\mathrm{CO}}_2}^{\mathrm{skin}}}\left[P_{{\mathrm{CO}}_2}^{\mathrm{tc}}\left(T^{10}\right)-P_{{\mathrm{CO}}_2}^{\mathrm{amb}}\left(T^{10}\right)\right]& \left(59\right)\end{array}$

[0195] D.sub.CO.sub.2.sup.skin is a volumetric flow rate of diffusion of CO.sub.2 through the skin, such that:

D.sub.CO.sub.2.sup.skin=ρ.sub.skin.sup.blood.Math.S.sub.skin.Math.D.sub.skin.sup.bin  (60)

[0196] D.sub.air(T.sup.10) is a volumetric flow rate of ambient air in the collecting chamber 10, such that:

[00038]$\begin{array}{cc}{D}_{\mathrm{air}}\left(T^{10}\right)=R.Math.T^{10}\frac{\left(P_{\mathrm{air}}^{\mathrm{out}\mathrm{col}}\left(T^{10}\right)-P_{\mathrm{air}}^{\mathrm{amb}}\left(T^{\mathrm{amb}}\right)\right)}{P_{\mathrm{air}}^{\mathrm{out}\mathrm{col}}\left(T^{10}\right).Math.R^{\mathrm{col}-\mathrm{amb}}}& \left(61\right)\end{array}$

[0197] P.sub.CO.sub.2.sup.amb(T.sup.10) is the CO.sub.2 pressure, in the collecting chamber, due to the ambient air, such that:

[00039]$\begin{array}{cc}{P}_{{\mathrm{CO}}_2}^{\mathrm{amb}}\left(T^{10}\right)=\frac{T^{10}}{T^{\mathrm{amb}}}.Math.P_{{\mathrm{CO}}_2}^{\mathrm{amb}}\left(T^{\mathrm{amb}}\right)& \left(62\right)\\ P_{\mathrm{CO}2}^{\mathrm{blood}}=\frac{P_{\mathrm{CO}2}^{10}}{P_{\mathrm{air}}^{10}}\left[P_{\mathrm{air}}^{\mathrm{amb}}\frac{T^{10}}{T^{\mathrm{amb}}}\sqrt{\frac{S^{15}}{S^{11}}}+\frac{\left(P_{\mathrm{air}}^{10}-P_{\mathrm{air}}^{\mathrm{amb}}\right)}{R^{\mathrm{col}-\mathrm{amb}}{S}_{\mathrm{skin}}{\sigma }_{\mathrm{CO}2}D}\right]& \left(63\right)\end{array}$

[0198] Taking into account the Severinghaus model:

[00040]$\begin{array}{cc}{P}_{{\mathrm{CO}}_2}^{\mathrm{blood}}\left(37°C.\right)=\frac{P_{{\mathrm{CO}}_2}^{\mathrm{blood}}\left(T^{10}\right)}{\exp \left(0.046\left[T^{10}-37\right]\right)}& \left(64\right)\end{array}$

[0199] In the preceding expression, T.sup.10 is expressed in ° C.

[0200] Step 90: Calibrating the Model

[0201] Step 100 requires the model parameters m.sub.1, n.sub.1, m.sub.2, n.sub.2, et q and the values of the powers u.sub.1, u.sub.2 to have been provided. To do this, prior to the implementation of the method, a calibration is carried out with the aim of determining the values of the parameters and of the powers. The calibration may be a supervised calibration, using a number N.sub.cal of calibration samples the CO.sub.2 content of which is known.

[0202] The parameters and powers are estimated by minimizing a functional J, such

[00041]$\begin{array}{cc}\mathrm{that}J=\underset{i=1}{\overset{N_{\mathrm{cal}}}{.Math.}}{\left[-\ln \left[\frac{\frac{U_i\left({\lambda }_1\right)}{U_i\left({\lambda }_2\right)}}{\frac{U_0\left({\lambda }_1\right)}{U_0\left({\lambda }_2\right)}}\right]-A_i\right]}^2\mathrm{with}{A}_i=m_1{C}_{i_{\mathrm{CO}2}}^{u_1}+n_1{C}_{i_{\mathrm{CO}2}}^{2u_1}+m_2{C}_{i_{H2O}}^{u_2}+n_2{C}_{i_{H2O}}^{2u_2}+{\mathrm{qC}}_{i_{\mathrm{CO}2}}^{u_1}.Math.C_{i_{H2O}}^{u_2}& \left(70\right)\end{array}$

[0203] In other words,

[00042]$\begin{array}{cc}\left\{m_1,n_1,m_2,n_2,q,u_1,u_2\right\}=\mathrm{argmin}J=\mathrm{argmin}\left(\underset{i=1}{\overset{N_{\mathrm{cal}}}{.Math.}}{\left[-\ln \left[\frac{\frac{U_i\left({\lambda }_1\right)}{U_i\left({\lambda }_2\right)}}{\frac{U_0\left({\lambda }_1\right)}{U_0\left({\lambda }_2\right)}}\right]-A_i\right]}^2\right)& \left(71\right)\end{array}$

[0204] Generally, it is a question of determining parameters such that, taking into account the known CO.sub.2 concentration, the application of the model approaches most closely the measured quantities.

[0205] Since the model is non-linear, the set of parameters {m.sub.1, n.sub.1, m.sub.2, n.sub.2, q, u.sub.1, u.sub.2} minimizing J may be determined by implementing a Levenberg-Marquardt algorithm.

[0206] A regularizing term may be introduced to make the estimation of the parameters more robust to measurement errors and to the approximations of the model:

{m.sub.1,n.sub.1,m.sub.2,n.sub.2,q,u.sub.1,u.sub.2}=argmin J

[00043]$\begin{array}{cc}=\mathrm{argmin}\left(\frac{1}{N_{\mathrm{cal}}}\underset{i=1}{\overset{N_{\mathrm{cal}}}{.Math.}}{\left[-\ln \left[\frac{\frac{U_i\left({\lambda }_1\right)}{U_i\left({\lambda }_2\right)}}{\frac{U_0\left({\lambda }_1\right)}{U_0\left({\lambda }_2\right)}}\right]-A_i\right]}^2+\omega \frac{1}{\dim \left(\left\{m_1,n_1,m_2,n_2,q,u_1,u_2\right\}\right)}{.Math.\left\{m_1,n_1,m_2,n_2,q,u_1,u_2\right\}.Math.}^2\right)& \left(72\right)\end{array}$

where:

[0207] dim({m.sub.1, n.sub.1, m.sub.2, n.sub.2, q, u.sub.1, u.sub.2}) is the dimension of the vector of parameters, which in this case is equal to 7

[0208] ∥{m.sub.1, n.sub.1, m.sub.2, n.sub.2, q, u.sub.1, u.sub.2}∥ is the norm of the vector of parameters. In the case of a quadratic norm:

∥{m.sub.1,n.sub.1,m.sub.2,n.sub.2,q,u.sub.1,u.sub.2}∥.sup.2=m.sub.1.sup.2+n.sub.1.sup.2+m.sub.2.sup.2+n.sub.2.sup.2+q.sup.2+u.sub.1.sup.2+u.sub.2.sup.2

[0209] The parameter ω may be chosen a priori, or determined automatically using techniques such as cross-validation.

[0210] To force the model to resemble a standard quadratic linear model, the powers may be required to be close to 1. In this case:

[00044]$\begin{array}{cc}\left\{m_1,n_1,m_2,n_2,q,u_1,u_2\right\}=\mathrm{argmin}J=\mathrm{argmin}\left(\frac{1}{N_{\mathrm{cal}}}\underset{i=1}{\overset{N_{\mathrm{cal}}}{.Math.}}{\left[-\ln \left[\frac{\frac{U_i\left({\lambda }_1\right)}{U_i\left({\lambda }_2\right)}}{\frac{U_0\left({\lambda }_1\right)}{U_0\left({\lambda }_2\right)}}\right]-A_i\right]}^2+\omega \frac{1}{\dim \left(\left\{m_1,n_{1-1},m_2,n_2-1,q,u_1,u_2\right\}\right)}{.Math.\left\{m_1,n_1-1,m_2,n_2-1,q,u_1,u_2\right\}.Math.}^2\right)& \left(73\right)\end{array}$

[0211] During training tests, the dilution correction factor g.sub.CO.sub.2.sup.tc/col may be estimated by carrying out a plurality of training acquisitions on volunteers for whom reference measurements, either of transcutaneous pressure or of blood pressure, are carried out at the same time as the measurement of the device. The average value of the factor g.sub.CO.sub.2.sup.tc/col to be applied so that the estimated values of transcutaneous pressure P.sub.CO.sub.2.sup.tc(T.sup.tc) approach as closely as possible the measured reference transcutaneous-pressure values, or so that the estimated values of blood pressure P.sub.CO.sub.2.sup.blood (37° C.) approach as closely as possible the measured reference blood-pressure values, is then estimated.

[0212] Example of Gas-Phase Calibration

[0213] In a first example, the nonlinear model described in expression (3) was calibrated, in step 90, by feeding the measuring chamber with a gas the CO.sub.2 concentration of which described, in each trial, concentration plateaux equal to 0, 0.2, 0.4, 0.8, 1.6 and 3.2 mol/m.sup.3, respectively. 12 trials were carried out, while varying the temperature T.sup.20 of the measuring chamber 20, the water concentration value C.sub.H2O, and the temperature T.sup.10 in the collecting chamber 10.

[0214] Table 1 shows the experimental parameters taken into account in each trial.

TABLE-US-00001 TABLE 1 Trial C.sub.H2O (mol/m.sup.3) T.sup.10 (° C.) T.sup.20 (° C.) 1 0.126 30.1 30.3 2 0.301 39.7 32.0 3 0.365 37 32.4 4 0.364 39.7 31.8 5 0.691 40.1 32.2 6 0.428 30.2 30.2 7 0.623 37.1 32.0 8 0.898 40.1 31.5 9 0.863 37.2 31.6 10 0.609 30.3 30.3 11 0.753 30.1 30.2 12 1.160 37.1 32.3

[0215] The parameters and powers of the model were estimated using three models:

[0216] model A: without taking into account water vapor, such as described with reference to expression (3′);

[0217] model B: model of expression (3) without taking into account the bilinear term, this amounting to setting q=0;

[0218] model C: model of expression (3).

[0219] Table 2 collates the model parameters obtained following the minimization described in step 90.

TABLE-US-00002 TABLE 2 Parameters Model A Model B Model C ln( ) 1.08 0.99 1.07 m.sub.1 0.29 0.29 0.29 u.sub.1 0.61 0.59 0.57 n.sub.1 0.051 0.048 0.044 m.sub.2 0.0040 0.0015 u.sub.2 0.12 2.45 n.sub.2 0.065 0.00001 q 0.00028

[0220] In table 2, the parameter ln( ) corresponds to

[00045]$\ln \left[\frac{U_0\left({\lambda }_1\right)}{U_0\left({\lambda }_2\right)}\right].$

This value was measured before carrying out the trials, in the absence of CO.sub.2 from the measuring chamber.

[0221] The three models were compared by comparing, during each trial, and for each CO.sub.2 concentration, the measured values with the actual values of the concentrations. For each model, a root-mean-square error (RMSE) and the coefficient of determination (R.sup.2) were calculated. Table 3 collates the obtained results.

TABLE-US-00003 TABLE 3 Model RMSE R.sup.2 A 0.0180 0.982 B 0.0140 0.989 C 0.0133 0.990

[0222] The results collated in table 3 show that model C, such as described with reference to expression (3), has the lowest RMSE and the highest coefficient of determination. It is therefore optimal. It may also be seen that taking into account water-vapor concentration (models B and C) allows performance to be improved.

[0223] FIG. 7A shows, for each trial, the exact measured values of

[00046]$-\ln \left[\frac{U\left({\lambda }_1\right)}{U\left({\lambda }_2\right)}\right]$

(crosses) and the values of

[00047]$-\ln \left[\frac{U\left({\lambda }_1\right)}{U\left({\lambda }_2\right)}\right]$

calculated implementing model C (circles). The consistency between the measured and calculated values may be seen, this attesting to the appositeness of the model.

Liquid-Phase Calibration

[0224] In order to estimate the constants of the fluidic model, it is also possible to carry out a calibration using mixtures in a reference liquid such as water, for known CO.sub.2 concentrations. The CO.sub.2 dilution required to achieve the mixtures may be accelerated using a bubbler and a mixer. Various techniques allow the concentration of CO.sub.2 dissolved in a liquid, water for example, to be controlled, for example use of a dedicated probe, for example a submersible probe “Membrane/NDIR CO.sub.2 probe” sold by AquaMS. The concentration of dissolved CO.sub.2 may also be obtained by measuring pH (hydrogen-ion content) and kH (carbonate-ion content). The carbon dioxide released from the mixture is measured with the device such as described above. The concentration of carbon dioxide dissolved in the water is estimated. The model parameters are estimated so as to minimize the error between the estimated carbon-dioxide concentration and the physical concentrations of the standard mixtures.

Training-Based Calibration on Volunteers

[0225] In order to estimate the parameters related to the properties of the skin and of the blood, and also to calibrate the parameters of the models, it is possible to carry out training measurements on volunteers for whom reference measurements are also carried out. These reference measurements may be assays on blood samples, or measurements of exhaled air, or a measurement of transcutaneous pressure using a commercial device such as an electrochemical measuring device. The coefficients of the model that are not measured by sensors are estimated by taking an average value that will be applied to all of the training measurements. These parameters are estimated by minimizing the quadratic deviation between the pressures estimated by the model on the basis of the measurements carried out by a device 1 as described above and the pressures delivered by the reference devices.

Trial on a User

[0226] The device 1 was applied against the forearm at the elbow crease of a healthy user undergoing a cardiopulmonary exercise test. In this test, the user was placed on a bicycle and the effort required from him modulated via pedaling speed. After a period of rest, lasting 9 minutes, 4 stages of cycling for 5 minutes were undergone, the pedaling speed gradually being increased. Two recovery stages were then administered: one at slow speed, the other at rest. FIG. 7B shows a variation in the partial pressure of CO.sub.2 P.sub.CO2.sup.10 in the collecting chamber, as measured by implementing steps 100 to 120 described above. The measurement points have been represented by dark crosses. The y-axis corresponds to the estimated partial pressure, whereas the x-axis corresponds to time. Given the magnitude of the noise, relative to the dynamic range of the signal, the measured values were smoothed either via a moving average of 100 s width (curve a), or via a polynomial of order 3 (curve b). The smoothed values are consistent with the variation in the partial pressure of CO.sub.2 in the blood during cardiopulmonary exercise according to such a protocol.