Non-invasive sensor and measurement method

Abstract

The present invention relates to a method for measuring a parameter of interest in a target layered medium (2) by means of a non-invasive sensor (1) based on photoacoustic detection or photothermal detection. The method comprises: a) providing a sensor comprising a light source (11a), a device for controlling a plurality of irradiation parameters of the light source (11a), a detection cell (12), a memory storing a nomogram of nodal frequencies, and an adaptation module (14); b) the adaptation module (14) selecting an initial model configuration (CMirrad) of the target layered medium for a given irradiation; c) a processor of the adaptation module (14) determining, in the nomogram of nodal frequencies, a plurality of nodal modulation frequencies, and then at least one specific modulation frequency (foptim), and finally a specific irradiation case including at least one specific modulation frequency (foptim); d) the light source (11a) irradiating the target layered medium (2) according to the specific irradiation case; e) the detection cell (12) detecting an acoustic signal or a thermal signal generated in response to the irradiation; f) a processor of the adaptation module (14) determining the parameter of interest based on the detected acoustic signal or thermal signal.

Description

Technical Field

[0001] The present invention relates to process measurement and non-invasive sensors that enable the measurement of one or more parameters of interest in a target environment.

[0002] More precisely, the present invention relates to a photo-invasive sensor based on the detection of a photo-thermal effect or photoacoustics, configured to measure one or more parameters in a target layered environment where layering can develop over time. The parameter to be measured can be, for example, blood glucose in the skin.

Background Art

[0003] In the field of sensors for living organisms, it is known to implement photo-invasive sensors based on photoacoustic or photo-thermal detection.

[0004] A region of interest in the environment to be analyzed, called the target, is irradiated by a laser beam of a wavelength and modulation frequency selected depending on the parameter of interest to be measured. The laser beam is absorbed by the target over a representative length depending on the structure of the target. The energy-absorbing light causes local heating of the target. In response to this heating, a heat wave of a frequency equal to the laser modulation frequency is generated in the target. This wave propagates within the target, in particular up to the outer surface of the target.

[0005] The heat wave can be directly detected and analyzed. Next, the photo-thermal effect will be described. Photoacoustic detection utilizes the fact that the heat wave is associated with a pressure wave frequency identical to the modulation frequency.

[0006] In the case of indirect photoacoustic detection, when the heat wave generated in the target reaches the target-external fluid environment interface after propagation, a pressure wave generated in the external fluid environment is detected.

[0007] The photoacoustic effect has been the subject of numerous theoretical studies. Allan Rosencwaig and Allen Gersho significantly developed a theoretical model for photoacoustic signals. This model involves the physicochemical properties of the sample being analyzed, including the optical diffusion length, the thickness of the sample, and the thermal diffusion length. (Rosencwaig, A. and Gersho, A. (1976), Theory of the photoacoustic effect with solids, Journal of Applied Physics, 47, 64).

[0008] Hu et al. developed a theoretically generalized photoacoustic effect in layered materials. (Hu, H., Wang, X., & Xu, X. (1999). Generalized theory of the photoacoustic effect in a multilayer material. Journal of Applied Physics, 86, 3953 - 3958).

[0009] Photoacoustic detection presents many advantages compared to other detection techniques, among which is the aspect of the orthogonality of transduction, i.e., the optical signal in the input of the environment being analyzed is converted into an acoustic signal that is highly specific to the observed phenomenon, which enables the use of inexpensive and small sensors.

[0010] The difficulties of photoacoustic detection or photothermal detection include the following among others: The large number of parameters that generally affect the detected signal, and For certain analytes present at low concentrations in the environment being analyzed, the proportion of the detection signal specific to these parameters of interest is low.

[0011] Photoacoustic or photothermal detection requires not only the selection of the wavelength but also the selection of the laser modulation frequency. In fact, the characteristic penetration length of the incident light wave into the target depends on the laser modulation frequency. When the structure of the target (e.g., skin) changes over time, the modulation frequency used to measure the parameter of interest (e.g., blood glucose in interstitial fluid) with controlled accuracy and power consumption also changes over time.

[0012] Furthermore, in the target layered environment, due to the reflection at the interfaces between various layers, so-called "primary" thermal waves propagate towards the surface of the target environment, generating a number of so-called "secondary" and "tertiary" waves. As shown by many researchers, these waves are likely to interfere with each other. The interference of the thermal waves generated by the photothermal effect is described in the following documents: C. A. Bennett and R. R. Patty, “Thermal wave interferometry: a potential application of the photoacoustic effect,” Appl. Opt. 21, 49-54(1982); Andreas Mandelis, “Theory of photothermal-wave diffraction and Interference in condensed media,” J. Opt. Soc. Am. A 6, 298-308(1989); Andreas Mandelis and Kwan F. Leung, “Photothermal-wave diffraction and interference in condensed media: experimental evidence in aluminum,” J. Opt. Soc. Am. A 8, 186-200(1991).

[0013] These interferences are likely to affect the accuracy of non-invasive sensors based on photoacoustic or optothermal detection. In a target layered environment where layering is likely to develop, the interference conditions also depend on the structuring of the target environment and are therefore likely to develop. As a result, the interference conditions are not constant over time.

[0014] Accordingly, the present invention aims to improve the accuracy of non-invasive sensors in a target environment, particularly in a layered and / or evolving environment, based on photoacoustic or optothermal detection, by taking into account interference phenomena while controlling the energy consumption of the sensor, or to reduce the energy consumption of the sensor while controlling its accuracy.

Summary of the Invention

[0015] For this reason, the present invention relates to a method for measuring a parameter of interest in a target environment means of a non-invasive sensor based on photoacoustic detection or optothermal detection. The measurement method includes the following: a) providing a sensor, the sensor comprising: a light source, a device for controlling a plurality of irradiation parameters of the light source, the plurality of irradiation parameters including at least one frequency for modulating the intensity of the light source, a detection cell (12) configured to detect an acoustic signal or a thermal signal, a memory storing a node frequency chart, the memory comprising: a plurality of model configuration groups of the target layered environment, each group including at least two model configurations of the target layered environment in which only the parameter of interest differs from each other, and A plurality of irradiation case groups, each including at least two irradiation cases that differ from each other only in the modulation frequency of the light source, and each irradiation case includes a set of irradiation parameter values. For the plurality of irradiation case groups, it includes a plurality of node multiplets, and each node multiplet includes characteristics common to all elements of a given model configuration group of the target layered environment and a plurality of associated node modulation frequencies. For the plurality of associated node modulation frequencies, the acoustic signal or thermal signal detected by the detection cell in response to irradiation by the light source exhibits a correlation with the parameter of interest that is less than a predetermined threshold value. A memory; Comprising a processor, and an adaptation module that exchanges information with the detection cell and the light source irradiation parameter control device A sensor is provided; b) The adaptation module selects a model configuration (CMirrad) of the initial target layered environment for irradiation; c) The adaptation module processor: In the node frequency chart, determine a plurality of node modulation frequencies associated with the model configuration group of the target layered environment to which the selected model configuration of the layered environment belongs, Next, based on these plurality of node modulation frequencies, determine at least one specific modulation frequency, which is a weighted average of at least two of these plurality of node modulation frequencies, bounded by two node modulation frequencies, Determine a specific irradiation case of the selected model configuration of the target layered environment, and the specific irradiation case includes at least one specific modulation frequency; d) The light source irradiates the target layered environment according to the set of irradiation parameters of this specific irradiation case; e) The detection cell detects the acoustic signal or thermal signal generated in response to the irradiation; f) The processor of the adaptation module determines the parameter of interest based on the detected acoustic signal or thermal signal.

[0016] With these provisions, the irradiation is carried out according to an irradiation case that includes the modulation frequency of at least one light source, especially in that the irradiation is selected to take into account the node frequencies corresponding to the model configuration of the target layer environment selected for the irradiation. In particular, a specific modulation frequency is determined based on two node frequencies, such that the specific frequency can correspond to the maximum sensor sensitivity for the interval between these two modulation frequencies (when all other things are equal).

[0017] Thus, these provisions make it possible to improve the sensitivity of non-invasive sensors compared to prior art sensors that do not include this step of adapting the irradiation case by taking into account interference phenomena.

[0018] Furthermore, it is possible to select a limited number of frequencies for the irradiation, for example, by retaining only a specific modulation frequency, or only a small number of modulation frequencies that include this specific modulation frequency.

[0019] Determining a specific modulation frequency constrained by the node frequencies as a weighted average of the node frequencies takes into account that the modulation frequency corresponding to the maximum sensitivity to the parameter of interest is within the interval between two consecutive node modulation frequencies, but the characteristic penetration depth of the incident light decreases with the modulation frequency. Thus, in order to improve the accuracy of the non-invasive sensor, it may be advantageous to select a weighted average of these node modulation frequencies by not selecting the arithmetic mean of the two node modulation frequencies as the specific modulation frequency, and in particular, by assigning a higher weight to the lowest node modulation frequency.

[0020] Thus, these provisions make it possible to reduce the irradiation power while maintaining the same sensitivity as that of prior art sensors.

[0021] According to one embodiment of the measurement method, The processor of the adaptation module further implements an inverse modeling algorithm that receives, as input, an irradiation case and an acoustic or thermal signal, and provides, as output, a model configuration of the target layered environment and values of the parameters of interest. Step f) includes the following steps: f1) The adaptation module processor receives, as input, an acoustic or thermal signal detected by a detection cell and a specific irradiation case used for irradiation, and returns, as output, from the inverse modeling algorithm, an estimated value of the model configuration of the current target layered environment and the parameters of interest; f2) The adaptation module processor evaluates the model configuration of the selected target layered environment by comparison with the model configuration of the current target layered environment, and only if this evaluation is unfavorable: g) The adaptation module receives, as input, the model configuration of the current target layered environment and returns, as output, a new model configuration of the target layered environment selected for irradiation, and then steps c), d), e) and f1) are repeated; f3) The value of the parameter of interest measured by the sensor is the last value of the estimated parameter of interest.

[0022] The evaluation of the model configuration CMirrad of the target layered environment is considered unfavorable when the model configuration CMirrad of the target layered environment is different from the model configuration CMmes of the target layered environment.

[0023] With these provisions, it is possible to obtain a blood glucose measurement using: A single irradiation (in the case of a favorable evaluation, sub-step g is not executed). This irradiation is then performed according to an irradiation case that is selected by default but corresponds to a controlled energy consumption (the irradiation parameters are selected, for example, from a limited number). The step f2) of evaluating the model irradiation configuration ensures that the measurement accuracy is as high as possible for the given irradiation case available in the correspondence table. This is advantageous over a process where step g) is not provided and this certainty is not available; Or step f2) involving the execution of two irradiations (sub-step g in the case of an undesired evaluation), and thus including the re-iteration of c), d), e) and f1). Next, the second irradiation is carried out according to the irradiation case that achieves the maximum possible accuracy for the case in terms of the additional information obtained from the first irradiation, and the irradiation cases available in the correspondence table. Thus, the accuracy of the process is known and improved compared to the process without step g). Furthermore, the selected irradiation case not only takes into account the interference phenomena for the model configuration selected for the first irradiation, but also verifies the optimal characteristics of the model configuration with respect to the current state of the layered environment, and if necessary, another model configuration and another irradiation case taking into account the interference phenomena are selected for the second irradiation based on which interest parameters are determined.

[0024] In this way, the accuracy and / or sensitivity and / or energy consumption of the non-invasive sensor can be continuously improved along with the development of the target layered environment.

[0025] According to one embodiment of the measurement process, g1) further includes re-iterating f2) at the end of f1).

[0026] According to this configuration, taking into account the interference phenomenon, the model configuration and the selection of the irradiation case are repeatedly iterated until the desired sensor performance in terms of the accuracy and / or sensitivity and / or energy consumption of the non-invasive sensor is achieved. For this reason, the inventors do not limit to one or two irradiations, and three or more irradiations may ultimately achieve better sensitivity to the situation of the target layered environment at the measurement time point than one or two irradiations, or may achieve a better compromise of sensitivity / accuracy / energy consumption. Here also, the energy consumption for each irradiation is controlled by taking into account the interference phenomenon even when the number of irradiation parameters is limited for each irradiation case and these irradiation parameters converge only after 3, 4, 5 or even 10 consecutive irradiations, and the total energy consumption can be controlled, in particular, to be lower than that required to collect the complete spectrum at the irradiation frequency and / or modulation frequency while controlling and even further improving the measurement accuracy at the same time.

[0027] According to one embodiment of the measurement process, the process preliminarily includes the following: I - A node frequency chart is generated by an I - processor, a database of model configurations including a multiplex (model configuration of the target layered environment, irradiation case, parameter of interest), and an acoustic signal or a thermal signal detected by a detection cell associated with each of these multiplexes, for: a plurality of model configuration groups, and a plurality of irradiation case groups and is stored in the memory of the non-invasive sensor. The node frequency chart includes a plurality of node multiplexes, and each node multiplex includes characteristics common to all elements of the model configuration group of the target layered environment and a plurality of associated node modulation frequencies. This node frequency chart is stored in the memory of the non-invasive sensor.

[0028] According to this configuration, the node frequency chart can be generated in a much shorter time by simulation, i.e., in silico, and covers a much more complete range of model configurations and irradiation cases than when the chart is generated manually. This configuration further improves the sensor accuracy. It also reduces the cost and manufacturing time of the non-invasive sensor.

[0029] According to one embodiment, the measurement method includes the following: II - The processor learns at least one inverse modeling algorithm from the configuration database model and stores the at least one inverse modeling algorithm in the memory of the non-invasive sensor.

[0030] With this configuration, for example, one or more inverse modeling algorithms, each adapted to a given model configuration and irradiation case or a group of given model configurations and irradiation cases, can be trained in an automated manner, thereby enabling the accuracy in the inverse problem-solving steps required to determine the parameter values of interest from the photoacoustic or photothermal signal detection signals.

[0031] According to one embodiment of the measurement process, at least some of the acoustic or thermal signals detected by the detection cells associated with the multiplex stored in the database of the model configuration are simulated, i.e., they are generated by a computerized simulation device.

[0032] According to this configuration, a large number of irradiation case groups and configuration groups can be generated in a much shorter time by simulation, i.e., in silico, and cover a much more complete range of model configurations and irradiation case groups than is accessible through in vivo experience. This configuration further improves the sensor accuracy. It also reduces the cost and manufacturing time of the non-invasive sensor.

[0033] According to one embodiment of the measurement method, the specific modulation frequency determined in step c) is a weighted average of at least two of the plurality of node modulation frequencies.

[0034] The present invention also relates to a non-invasive sensor based on photoacoustic or photothermal detection configured to measure a parameter of interest in a layered target environment. The non-invasive sensor comprises: a light source, a device for controlling the irradiation parameters of the light source, a detection cell configured to detect an acoustic signal or a thermal signal, a memory storing a node frequency chart, comprising: a plurality of model configuration groups of the target layered environment, each group comprising at least two model configurations of the target layered environment in which only the parameter of interest differs from each other, and a plurality of irradiation case groups, each group comprising at least two irradiation cases that differ from each other only in the modulation frequency of the light source, each irradiation case comprising a set of irradiation parameter values, for the plurality of irradiation case groups, comprising a plurality of so-called "node" multiplexes, each multiplex comprising a characteristic common to all elements of a given model configuration group of the target layered environment and a plurality of associated node modulation frequencies, for the plurality of associated node modulation frequencies, the acoustic signal or the thermal signal detected by the detection cell in response to irradiation by the light source exhibits a correlation with the parameter of interest below a predetermined threshold, the memory and the non-invasive sensor comprises a processor and further comprises an adaptation module adapted to exchange information with the detection cell and the light source irradiation parameter control device, the adaptation module: selects a model configuration of the initial target layered environment, determines a plurality of node modulation frequencies associated with the model configuration group to which the selected model configuration of the target layered environment belongs in the node frequency chart, Determining at least one specific modulation frequency based on these multiple node modulation frequencies, where this specific modulation frequency is a weighted average of at least two of these multiple node modulation frequencies and is bounded by two node modulation frequencies; Determining a specific irradiation case for a selected irradiation model configuration including at least one specific modulation frequency; Transmitting the specific irradiation case to a light source irradiation parameter control device; Receiving an acoustic signal or a thermal signal detected by a detection cell; Determining a value of a parameter of interest based on the detected acoustic signal or thermal signal; It is further configured to do so.

[0035] In one embodiment of the non-invasive sensor, the processor of the adaptation module: Implementing an inverse modeling algorithm that receives, as input, an irradiation case including a set of irradiation parameters and an acoustic signal or a thermal signal and outputs a model configuration of the target layered environment and a value of a parameter of interest; Determining, by the inverse modeling algorithm, a model configuration of the current target layered environment and a value of a parameter of interest estimated based on the received and detected photoacoustic signal or photothermal signal and the selected irradiation case; Evaluating a selected model irradiation configuration of the target layered environment by comparison with the model configuration of the current target layered environment; Only if the selected irradiation model configuration of the target layered environment is different from the model configuration of the current target layered environment being compared: when receiving the model configuration of the current target layered environment, determining a new model configuration of the selected target layered environment for irradiation; Determining, in a node frequency chart, a plurality of node modulation frequencies associated with a model configuration group of the target layered environment to which the selected new model configuration of the layered environment belongs; Next, based on these multiple node modulation frequencies, at least one specific modulation frequency is determined, and this specific modulation frequency is a weighted average of at least two of these multiple node modulation frequencies, and the boundary is defined by two node modulation frequencies. Determine a new specific irradiation case of the newly selected model configuration of the target layered environment, and the specific irradiation case includes at least one specific modulation frequency. The light source irradiates the target layered environment according to a set of irradiation parameters of this new specific irradiation case, the detection cell detects an acoustic signal or a thermal signal generated in response to this new irradiation, and the processor of the adaptation module determines an interest parameter based on the detected new acoustic signal or thermal signal, and then transmits the new specific irradiation case to the light source irradiation parameter control device; Determine the value of the interest parameter measured based on the last value of the estimated interest parameter. It is further adapted to do so.

[0036] According to an embodiment of the non-invasive sensor, at least two of the light source, the device for controlling the irradiation parameters of the light source, the detection cell configured to detect an acoustic signal or a thermal signal, the memory storing the node frequency chart, and the adaptation module are mechanically independent of each other.

[0037] By this type of implementation, it is possible to form a distributed sensor such that the space requirements or weight are reduced at the level of the target layered environment. In particular, the simulation module can be placed externally (deport) for remote management and, in some cases, can be shared among several users. Finally, the present invention relates to a computer program including instructions for causing a non-invasive sensor to execute the steps of the measurement process according to one of the embodiments described above according to one of the preceding embodiments.

[0038] Embodiments of the present invention will be described below with reference to the drawings briefly described below.

Brief Description of the Drawings

[0039]

Figure 1

Figure 2

Figure 3a

Figure 3b

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8a

Figure 8b

Figure 8c

Embodiments for Carrying Out the Invention

[0040] The present invention relates to a non-invasive sensor for one or more parameters of a target layered environment 2, the structure of which may change over time in some cases. The target layered environment 2 can be, for example, a tissue such as the skin of a human or animal body. The parameter to be measured is hereinafter referred to as the "parameter of interest".

[0041] In particular, the parameter of interest can be a physiological parameter where the layered environment is a human or animal tissue.

[0042] For example, the physiological parameter to be measured is blood glucose, in particular, blood glucose in interstitial fluid. This can also be a matter of measuring the water content in a specific layer of the skin, or the lactate concentration in a specific layer. These examples are not limiting.

[0043] The non-invasive sensor 1 can be made portable and can enable continuous monitoring of the parameter of interest.

[0044] The non-invasive sensor 1 can be based on photoacoustic detection or photothermal detection. The measurement method is particularly suitable for improving the accuracy of the non-invasive sensor based on indirect photoacoustic detection in which the non-invasive sensor 1 detects acoustic waves generated in the fluid surrounding the target layer environment 2 in response to irradiation while controlling energy consumption, or for reducing the energy consumption of the sensor while controlling the accuracy. However, it is also completely possible to implement this process in a non-invasive photothermal-based sensor.

[0045] For the sake of simplicity of understanding, an example of indirect photoacoustic detection is described in more detail below, but it can be easily generalized to photothermal-based sensors.

[0046] The non-invasive sensor 1 is schematically shown in FIG. 1, and the non-invasive sensor 1 includes: An irradiation device 11 including a light source 11a, an intensity modulation device 11b for this light source, and a control device 11c for modulating the intensity of the light emitted by the light source 11a at at least one frequency; At least one photoacoustic detection cell 12 for detecting a signal generated in response to irradiation of the target layer environment 2 by the emitted light, for example, directly or indirectly detecting a heat wave propagating in the target layer environment 2; A signal processing module 13 configured to receive and process data from at least one detection cell 12; An irradiation parameter and calibration model adaptation module 14; In some embodiments, a simulation module 15 that is remote from or embedded in other elements of the non-invasive sensor 1, is provided.

[0047] In certain embodiments, the light source 11a emits an intensity-modulated laser beam at at least one specific wavelength towards the target layered environment 2.

[0048] At least one wavelength can be selected as a function of one of the parameters of interest. By way of example, the wave number 1034 cm corresponding to the glucose absorption peak -1 may be relevant if the non-invasive sensor 1 is a blood glucose sensor.

[0049] In particular, the light source 11a can be a light-emitting diode (LED) or a laser chip. Alternatively, the light source 11a can include a quantum cascade laser (QCL) emitted in the mid-infrared region (MIR-QCL), an ICL laser (“intermediate band cavity laser”), an internal or external cavity laser, a GaSb laser. These examples are not limiting. The light source 11a can be selected according to the target layered environment 2 and / or the parameter of interest.

[0050] The non-invasive sensor 1 can comprise several different and / or identical light sources 11a.

[0051] The non-invasive sensor 1 also comprises a circuit associated with the light source 11a and at least one control device 11c configured to control the frequency at which at least one intensity modulation device 11b modulates the intensity of at least one light source 11a, whereby the intensity of the light emitted by the light source 11a is modulated by an adjustable modulation frequency.

[0052] The intensity modulation frequency of a given light source 11a is hereinafter referred to as fmod.

[0053] The light source 11a can be intensity-modulated by any known electrical or mechanical means.

[0054] The light irradiating the target layered environment 2 can be emitted continuously or in pulses by a given light source 11a.

[0055] The light incident on the target layered environment 2 emitted by the intensity-modulated light source 11a propagates into the target layered environment 2 and then passes through the target layered environment 2 (symbolized by the solid arrows in FIG. 1). Next, this is gradually absorbed over a characteristic depth represented by zmax, depending on the structure and its physicochemical composition of the target layered environment 2, by various components of this target layered environment 2.

[0056] The absorption of light energy causes local heating of the target layered environment 2. As a result, a heat wave having a frequency equal to the modulation frequency of the light source propagates through the target layered environment 2 (symbolized by the dotted arrows in FIG. 1), particularly towards the surface of the target layered environment 2. This heat wave generates a pressure wave of the same frequency in the gaseous environment outside the target, which propagates through this fluid environment, for example gaseous, surrounding the target layered environment 2, particularly within the photoacoustic detection cell 12 (the phenomenon symbolized by the alternating dotted arrows in FIG. 1). At this stage, the interference phenomenon of interest to the inventors is not described. This will be described later.

[0057] In the case of photoacoustic detection, the detection cell 12 includes a chamber filled with a gas (e.g., air) through which acoustic waves propagate. The detection cell 12 also includes one or more suitable sensors arranged within this chamber, for example facing the target layered environment 2. For example, one or more electroacoustic sensors configured to convert the pressure of the acoustic wave into an electrical signal, such as a microphone or a piezoelectric transducer.

[0058] Each electroacoustic sensor is functionally connected to the signal processing module 13.

[0059] The signal processing module 13 can include an analog / digital converter configured to convert an analog electrical signal from the electroacoustic sensor into a digital signal.

[0060] The signal processing module 13 can include a synchronous detection device adapted to modulate and extract a signal of interest from the detected signal.

[0061] Optionally, the signal processing module 13 includes an operational amplifier operatively connected to the analog / digital converter and configured to amplify an electrical signal derived from the acoustic response of the target layered environment 2 transmitted by the electroacoustic sensor.

[0062] In an exemplary embodiment, the analog / digital converter is operatively connected to a digital signal processor for digital signal processing.

[0063] The non-invasive sensor 1 according to the present invention further comprises an adaptation module 14 for the irradiation parameters, optionally a calibration model, and a simulation module 15. These two elements will be described after identifying the starting point of the technical problem to be solved, namely, providing the non-invasive sensor 1 based on photoacoustic or photothermal detection with controlled accuracy and / or consumption by taking into account the interference phenomenon.

[0064] Thermal wave interference phenomenon Since the target layered environment 2 includes at least two overlapping layers having different physicochemical properties, it has at least one interface capable of reflecting the incident wave.

[0065] In the simple case shown in FIG. 2, where the target layered environment 2 includes two layers 2A and 2B, the upper layer 2A is in contact with the external environment at the 2A / ext interface and with the lower layer at the 2A / 2B interface.

[0066] In the case of the non-invasive sensor 1 adapted for blood glucose measurement, the upper layer 2A is, for example, the stratum corneum of the patient's skin, and the lower layer 2B is, for example, the epidermis of the patient's skin.

[0067] Even in such a simple case, a detailed description of the thermal wave generated in the target layered environment 2 requires the use of a complex model such as the so-called "Hu model" described in Hu, Hanping et al. “Generalized theory of the photoacoustic effect in a multilayer material.” Journal of Applied Physics 86(1999):3953-3958. This model implicitly includes some thermal interference effects but does not take into account the interference caused by multiple reflections at the interfaces. Other models propose taking these multiple reflections into account, especially those described in Cao, J. (2000); Interferential formulization and interpretation of the photoacoustic effect in multilayered cells; Journal of Physics D: Applied Physics, 33(3), 200.

[0068] Regardless of the selected model, the interference phenomenon has a significant impact on the amplitude of the detected signal and, as a result, on the measurement accuracy. This impact depends on the structure of the target layered environment 2.

[0069] A rough understanding of this influence can be obtained from Figure 2 for educational purposes only. In this case, the so-called "primary" heat wave P is generated by irradiation at the interface of 2A / 2B (symbolized by the dotted arrow). This primary wave P propagates towards the interface of 2A / ext, where it is partially reflected, generating a first wave transmitted to the external environment T1 and a reflected wave R1. The reflected wave R1, in turn, can be reflected at the interface of 2A / 2B, generating a transmitted wave T2 and a reflected wave R2. The reflected wave R2, after propagating towards the interface of 2A / ext, generates a second wave transmitted into the external environment T2 and a reflected wave R3, etc.

[0070] All the waves transmitted to the external environments T1, T2,... are likely to interfere with each other, thereby affecting the amplitude of the acoustic waves formed in the external environment in response to the irradiation. In this case, it can be shown that the amplitude of the heat wave at the interface of 2A / air has the following formula:

Number

[0071]

Number

Number

[0072]

Number

[0073] The heat penetration rate and thickness of layer 2A vary over time in the case of the target layered environment 2 according to the present invention. In particular, when the target layered environment 2 is the skin and the non-invasive sensor 1 is a blood glucose sensor, the thickness of the skin layer, their water content, and glucose concentration, which are likely to affect different heat penetration rates, change over time, thereby affecting the photoacoustic or photothermal signal at the same detection site.

[0074] Therefore, it is understood that the accuracy and sensitivity of, for example, blood glucose measurement of the parameter of interest based on the photoacoustic and photothermal signals rely on the information available regarding these interference phenomena.

[0075] Here, the difficulty of accessing this information in the case of the target layered environment 2 under development is emphasized. Here, the characteristics of different layers change over time. To solve this problem, the inventors have developed a simulation module 15 that simulates a photothermal signal associated with irradiation with known characteristics (such as power, wavelength, modulation frequency, etc.), which is described in detail below. The target layered environment 2 is also modeled by a multi-layer environment with known characteristics (such as thickness, composition, etc.) based on a multiphysics analysis model that implicitly takes into account at least the thermal interference phenomenon. This simulation module 15 enables large-scale in silico experiments to be performed, and some of the results are shown in FIGS. 3a and 3b.

[0076] In FIG. 3a, the target layered environment 2 is the skin modeled by a two-layer model (stratum corneum, epidermis). The thickness of the upper layer (modeling the stratum corneum) is equal to 14 mm, and the thickness of the lower layer (modeling the epidermis) is equal to 10 mm. The layer (modeling the epidermis) is considered infinite. The blood glucose is 0 in the upper layer. The upper layer contains 20% water, and the lower layer contains 60% water.

[0077] The skin model is an intensity modulated 1034 cm -1 The amplitude of the simulated thermal response wave generated in response to the irradiation is plotted on the coordinates as a function of modulation frequency for different underlying blood glucose levels (50 mg / L; 500 mg / L and 1000 mg / dL).

[0078] Contrary to expectations, the inventors observed the existence of so-called "nodal" modulation frequencies, in the vicinity of which the amplitude of the heat wave is independent of the parameter of interest, specifically blood glucose.

[0079] In the case of Fig. 3a, two nodal modulation frequencies can be observed, namely fnod1 = 950 Hz and fnod2 = 4,750 Hz, but also outside the range of modulation frequencies shown in Fig. 3a. These nodal modulation frequencies can be explained by the phenomenon of thermal wave interference within the target layered environment 2.

[0080] The nodal modulation frequencies correspond to response waves generated in an environment whose amplitude is weakly correlated to the parameter of interest, or even independent of the parameter of interest (in this case, blood glucose in the interstitial fluid), and therefore the signal detected at such modulation frequencies cannot be used to determine the parameter of interest. In the case of FIG. 3b, the target layered environment 2 is a skin modeled by a two-layer model (stratum corneum, epidermis). The thickness of the upper layer (modeling the stratum corneum) is now equal to 18 mm, and the thickness of the lower layer (modeling the epidermis) is considered infinite. Blood glucose is 0 in the upper layer. The upper layer contains 20% water and the lower layer contains 60% water.

[0081] The skin model is an intensity modulated 1034 cm sigma laser with the modulation frequency plotted on the x-axis. -1 The amplitude of the simulated thermal response wave generated in response to the irradiation is plotted on the coordinates as a function of modulation frequency for different underlying blood glucose levels (50 mg / L; 500 mg / L and 1000 mg / L).

[0082] Here too, the node modulation frequencies were observed, and in this case, fnod1 = 600 Hz and fnod2 = 3000 Hz, but these node modulation frequencies are different from those obtained in the situation of Fig. 3a, as expected due to the changes in the thickness of layer 2A from 14 mm to 18 mm.

[0083] Generally speaking, the large-scale in silico experiments conducted by the inventors thus enabled them to observe the following facts: There are node modulation frequencies, in the vicinity of which the amplitude of the heat wave generated in response to irradiation is weakly correlated with the parameter of interest or even independent of the parameter of interest. When the scales of Figs. 3a and 3b are enlarged, it should be noted that this is not a problem of a single node frequency, but a problem of a limited range of frequencies where the heat amplitude hardly varies when the parameter of interest fluctuates; The values of the node modulation frequencies depend on the characteristics of the target layered environment 2, i.e., the thicknesses and compositions of the various layers that make it up.

[0084] Performing sufficient in vivo observations to obtain such curves is very difficult. Therefore, these observations were only made possible by the simulator developed by the inventors. Therefore, these observations enable the inventors to devise a new non-invasive sensor 1 that takes into account the interference phenomenon.

[0085] Based on these observations, the inventors designed a non-invasive sensor 1 and a method for detecting the parameter of interest that are improved over the prior art. This will be described in more detail below.

[0086] In the non-invasive sensor 1, due to the adaptation module 14 that characterizes it, and in a specific embodiment, the simulation module 15, the node modulation frequencies can be utilized in relation to the non-invasive sensor 1.

[0087] Adaptation module 14 The adaptation module 14 is a computerized device comprising at least one processor, which can: Exchange information with the simulation module 15 described below as appropriate; Receive information from the detection cell 12 and / or the signal processing module 13 as appropriate; Transmit information to the control device 11c of at least one irradiation parameter of the light source 11a (for example, the modulation frequency at which the intensity modulation device 11b modulates the intensity of the light emitted by the light source 11a, the wave number of the light emitted by the light source 11a, the light output of the light source 11a,...).

[0088] The steps carried out by the non-invasive sensor 1 comprising an adaptation module 14 for the measurement of the parameter of interest are schematically shown in FIG. 5 in a particular embodiment and in the case where the target layered environment 2 is the skin and the parameter of interest is the blood glucose in the interstitial fluid.

[0089] The non-invasive sensor 1 further comprises a memory for storing the database of the model configuration, a node frequency chart, and one or more inverse models described below.

[0090] This storage memory can be distributed in and / or shared with the adaptation module (14) and / or the simulation module (15).

[0091] The adaptation module (14) executes a process step for selecting the irradiation parameters for the measurement based on the node frequency chart and determining the most appropriate, i.e., the most accurate, inverse model for calculating the measured parameters from the signals detected for that particular measurement. These steps will be described after the steps for generating the database of the model configuration and the node frequency chart.

[0092] The model configuration database and the node frequency chart can be implanted in sensor 1 or generated from the remote simulation module 15. When the simulation module is remote, the non-invasive sensor 1 includes communication means for the simulation module 15 and an adaptation module 14 for exchanging data.

[0093] Simulation module 15 As shown in FIG. 6, the simulation module 15 is a computerized device configured to generate a set of model configurations, each corresponding (or alternatively modeling or describing) to a particular state of the photoacoustic signal (or, where appropriate, the optothermal signal) theoretically detected in response to each of a set of irradiation cases of the target layered environment 2, each of which includes at least one irradiation parameter such as light output, wavelength, and modulation frequency, and each model configuration CMk of the target layered environment 2 and each irradiation case of the photoacoustic (or, in some cases, optothermal) detection cell 12 from the analytical model of the target layered environment 2.

[0094] This simulation module 15 is particularly relevant in cases where the target layered environment 2 is evolving and / or stratifying over time. In this case, the target layered environment 2 adopts different actual configurations over time (one or more concentrations vary within one or more layers of the target layered environment 2, and one or more dimensions such as the thickness of one of the layers of the target layered environment 2 vary, for example), each of which can be modeled by a specific model configuration.

[0095] a) Multiphysics analytical model of the target layered environment 2 The layered target medium 2 to be analyzed is modeled as shown in FIG. 4. The target medium 2 is composed of N continuous layers that separate the outer medium A from the inner medium B, assuming that the interface is locally flat, for example.

[0096] Each layer [Number] is described by the parameter of interest and a plurality of explicit parameters suitable for the target layered environment 2 under consideration (when their values are provided as input to the simulation module 15 for each simulation, they are called level 1 parameters). For example, if the target layered environment 2 is the skin and the purpose is to measure the glucose concentration in layer j of this target environment 2, the target environment 2 is described by the number of layers N and the concentration of interest [Glc]j, and each layer i can be described for modeling by its thickness ei and its water concentration [H2O]i. The glucose concentration [Glc]i for i≠j can also be taken into account. In this case,

[0097] [Number] the value of, the [H2O]i concentration ( [Number] ) and the [Glc]i concentration (

[0098] [Number] ) are level 1 parameters.

[0099] The list of level 1 parameters can be extended if more precise modeling is desired. In particular, the concentrations of other skin components such as fat, lactic acid, oxygen, etc. can be included in the list of level 1 parameters that describe the skin layers.

[0100] Here again, for the example of the skin, it may be possible to take into account the skin color, the patient's age, or any other anthropometric parameter in order to expand or restrict the space of possible models.

[0101] The implicit parameters of the target layered environment model 2 (referred to as level 2 parameters because they are not provided as inputs to the simulation module 15) can be calculated by the analysis model. For example, the thermal conductivity, heat capacity, density, or absorption coefficient at each wavelength of each layer of the target layered environment 2 can be inferred from the level 1 parameters and known equations. Since the parameter of interest plays a special role, it is not included in the list of level 1 parameters. Depending on the process step, this parameter of interest may or may not be known, and its value is known for performing simulations using the simulation module 15, but of course is unknown in the case of actual measurements using the non-invasive sensor 1.

[0102] The number of layers N of the target layered environment 2 can also be model-variable. Again, in the example of the skin, depending on the physiological situation, N may be 2 or more. For this reason, for a particular physiological situation, the skin can be correctly described by two layers, namely, a first layer corresponding to the stratum corneum where the glucose concentration may be low, for example, and a second layer of the rest of the skin, and the glucose concentration of the second layer becomes the same (assimilated) as the glucose concentration in the measured interstitial fluid.

[0103] In the case of other situations, a model with three or more layers may be more appropriate. In the latter case, the moisture concentration of the layer may, for example, increase with the depth of the layer.

[0104] Therefore, N may not be constant.

[0105] In the case of the skin, the external medium A is usually the atmosphere surrounding the patient while also filling the photoacoustic detection cell.

[0106] The simulation module 15 performs a multi-physics analysis of the layered environment 2 model based on physical and / or chemical equations such as, for example, by way of non-limiting example, the Beer-Lambert equation for light absorption and the thermodynamic equations of heat (Fourier's law and conservation laws).

[0107] The model configuration CMk (where k is a positive integer) of the target layered environment 2 corresponds to (or models) a specific state of a given target layered environment 2. This specific state is assumed to be correctly represented by giving the number of layers N and the values of the level 1 parameters for each layer.

[0108] For each CMk model configuration, the multiphysics analysis model enables simulating the thermal wave generated at the interface 1 / A between layer 1 and the outer environment A in response to irradiation by a light source 11a for which the irradiation parameters, i.e., for example, the modulation frequency fmod, the wavelength l, and the power flux density, are known, when the parameter of interest is also known.

[0109] Alternatively, the multiphysics analysis model can be used to simulate the pressure wave generated in the external environment A.

[0110] In both cases, the signal obtained at the output of the processor implementing the multiphysics analysis model is called the "simulated response wave".

[0111] The simulated response wave can be provided as an input to the processor implementing the detection cell model.

[0112] b) Detection cell model The non-invasive sensor 1 based on photoacoustic or photothermal detection includes a photoacoustic detection cell 12 (or photothermal) configured to detect and analyze the pressure wave (or thermal wave) generated in the external environment A when the thermal wave generated in the target layered environment 2 reaches the interface 1 / A in response to irradiation.

[0113] The entire detection cell 12 can be analytically modeled. Theoretically, from the response wave simulated by the multiphysics analysis model, which would be received at the input of the detection cell 12, the model of the detection cell 12 predicts the output signal of the detection cell 12.

[0114] Various models can be devised.

[0115] For example, in the case of indirect photoacoustic detection, hereinafter, the parameters of the model of the detection cell 12, which are hereinafter referred to as cell parameters, can include its dimensions (i.e., vent size, cell height, etc.), thermodynamic state parameters (temperature, atmospheric pressure, relative or absolute humidity, etc.). In particular, the photoacoustic detection cell can be modeled by an equivalent RLC circuit. As an example, a model derived from the model described in Dehe, Alfons et al. “The Infineon Silicon MEMS Microphone.” (2013) may be suitable.

[0116] In this photoacoustic detection cell model, when applicable, in order to generate the signal theoretically obtained in the corresponding photoacoustic detection cell output (and after signal processing by the signal processing module 13 when applicable) from each simulated response wave generated by the processor that implements the analytical multiphysics model, it is possible to include a model of the signal processing steps executed by the signal processing module 13.

[0117] The processor of the simulation module 15 can be configured to implement the photoacoustic detection cell model.

[0118] Therefore, a global analysis model is available based on the multiphysics analysis model of the layered target environment 2 and the model of the detection cell 12. Assuming that the model configuration CMk of the layered target medium 2 and the irradiation case Ij are given and the parameter of interest is also supplied (since this is a simulation, it can be actually selected), the global analysis model can predict the signal expected at the output of the detection cell 12 or, as appropriate, the signal processing module 13. This is shown in FIG. 6.

[0119] The multiphysics analysis model is selected to at least implicitly account for thermal wave interference. In certain embodiments, these interferences can be explicitly accounted for.

[0120] Alternatively, all or some of the multiple reflections at the interface of the target layered environment 2 can be taken into account.

[0121] Thus, generally, the simulation module 15 receives as input the parameters of the model configuration CMk of the target layered environment 2, i.e., the number of layers N of the target environment 2, and the level 1 parameters of each layer, as well as the parameter of interest and the irradiation parameters of the irradiation case Ij. As output, the simulation module 15 provides a signal theoretically predicted at the output of the detection cell 12, or, as appropriate, a processed simulated signal theoretically expected at the output of the signal processing module 13 of the model configuration CMk of the target environment 2, also referred to as a simulated output signal.

[0122] The simulated output signal can be stored in memory in the form of a Fourier spectrum.

[0123] The multiplex {model configuration CMk of target environment 2, irradiation case Ij, parameter of interest, amplitude and phase of the simulated output signal component} can be stored in the model configuration database.

[0124] Database of model configurations Thus, a number of model configurations CMk can be generated, optionally automatically and / or randomly, each model configuration CMk corresponding to a set of the number of layers N and level 1 parameters, and optionally corresponding to a value or range of values of the parameter of interest that describes a particular situation of the target layered environment 2 of interest.

[0125] For each model configuration CMk, a large number of irradiation cases Ij can be generated, sometimes automatically and / or randomly, and each irradiation case corresponds to a set of irradiation parameters that describe the parameters of the light source 11a used for irradiation.

[0126] The irradiation case Ij can thus include the intensity of one or more lasers, the wavelength of each of these lasers, and optionally one or more frequencies for modulating the power irradiated by each laser.

[0127] At the output of the processor of the simulation module 15 that implements a global analysis model including a multiphysics analysis model and a detection cell model in a cascade connection by the simulation module 15, the amplitude and phase of each component of the simulated output signal are calculated for each model configuration CMk of each irradiation case Ij, and the values of the parameters of interest are also provided.

[0128] The irradiation case Ij may be the same for several different model configurations CMk and, in some cases, for several values of the parameter of interest, or may be different for each model configuration CMk and / or for each value of the parameter of interest. When the simulation is executed, all these model configurations CMk, irradiation cases, parameters of interest, and the associated simulated photoacoustic (or photothermal) signals can be stored in the model configuration database in the form of a multiplex {model configuration CMk, irradiation case Ij, parameter of interest, amplitude and phase of the components of the simulated output signal}.

[0129] The generation of the model configuration CMk and / or the irradiation case Ij may not be completely random.

[0130] The generation of the model configuration CMk can be based on physiological considerations among others in order to constrain the space of possibilities to physiologically realistic model configurations. For example, the possible thickness of the first skin layer can be restricted to the range [8 mm, 40 mm] that is actually experimentally observed, the water content of this layer can be restricted to the restricted range for each thickness, and the water concentration of the stratum corneum correlates with its thickness.

[0131] In particular, the generation of the irradiation cases can take into account the light source 11a available for a given non-invasive sensor 1 in terms of wavelength and / or power, or the modulation frequency range associated with the type of the target layered environment 2 to be analyzed, or the wavelength restrictions regarding the parameter of interest.

[0132] Alternatively, the database of model configurations may only contain the multiplets {model configuration CMk, irradiation case Ij, parameter of interest, amplitude and phase of the actually measured signal components} obtained in experiments in real-world situations, or may contain both such multiplets obtained in real-world situations and multiplets obtained by simulation.

[0133] Inverse model One or more artificial intelligence models can be trained in the simulation module 15 from the model configuration database.

[0134] As shown in FIG. 7, after learning, the artificial intelligence model solves the inverse problem, i.e., discovers the parameter of interest of the target environment 2 and the model configuration CMk, i.e., the number of layers N and the level 1 parameters, and can know the simulated photoacoustic or optothermal signal and the irradiation parameters of the irradiation case.

[0135] Thereafter, the learned model, hereinafter referred to as the inverse model, can be transmitted to the adaptation module 14 and stored in the memory of this model.

[0136] Several different inverse models can be learned with different training sets and / or training rules.

[0137] Node frequency chart A number of model configurations CMk can be generated, optionally and / or randomly, and each model configuration CMk corresponds to a set of the number of layers N and level 1 parameters, and optionally, it is already known that it corresponds to the value or range of values of the interest parameters that describe the specific situation of the target layered environment 2 of interest.

[0138] Therefore, a model configuration group GMp can be generated using two model configurations within the same group that differ only in the value of the interest parameter.

[0139] For each model configuration CMk, a number of irradiation cases Ij can be generated, optionally and / or randomly, and each irradiation case is also known to correspond to a set of irradiation parameters that describe the parameters of the light source 11a used for irradiation.

[0140] Therefore, it is also possible to generate an irradiation case group GIq, and two irradiation cases in each group GIq of irradiation cases differ only in the modulation frequency of the light source 11a.

[0141] For a group GMp of model configurations CMk that differ from each other only in the value of the interest parameter, the amplitude curve of the detection signal as a function of the modulation frequency of the light source can be generated by the irradiation simulated using a given irradiation case group GIq, and only the modulation frequency varies for each irradiation. In this way, FIGS. 3a and 3b were obtained.

[0142] From these curves, optionally, at least one node modulation frequency can be determined for this set GMp of model configurations by the processor of the simulation module 15 or the processor of the adaptation module 14.

[0143] Therefore, it is possible to determine a plurality of multiplets, hereinafter referred to as node multiplets. Each node multiplet includes the following: - Characteristics common to all elements of a given model configuration group GMp of the target layered environment (number of layers, layer thickness, concentration... excluding the parameter of interest that varies for each element of the GMp group), and - A plurality of node modulation frequencies associated with this GMp group. For the plurality of node modulation frequencies, the acoustic signal or thermal signal detected by the detection cell in response to irradiation by the light source exhibits a correlation less than a predetermined threshold with the parameter of interest.

[0144] Therefore, the node frequency chart can be generated from the database of model configurations, for example, using the simulation module 15.

[0145] When the simulation is executed, this node frequency chart can be stored in the memory of the simulation module. The chart includes at least one plurality of node multiplets, each of which includes at least characteristics (or a set of common irradiation parameters) common to the GMp model configuration group and the plurality of associated node modulation frequencies.

[0146] The node multiplet can optionally include one or more so-called specific frequencies denoted as foptim. For this frequency, the sensor sensitivity is maximum in the modulation frequency range bounded by two (optionally consecutive) node modulation frequencies of the parameter of interest.

[0147] Such a specific modulation frequency foptim can be observed in zone 1 shown in FIG. 3a.

[0148] In one embodiment, the specific modulation frequency is determined by the processor of the simulation module 15 or the adaptation module 14 based on the weighted average of at least two consecutive node modulation frequencies.

[0149] At this stage, the elements necessary for implementing the measurement process according to the present invention are ready.

[0150] It has already been understood that the node frequency chart allows taking into account the interference phenomenon, thereby reducing the number of irradiation modulation frequencies necessary to meet one or more optimization criteria, particularly among the energy consumption criterion, the sensitivity criterion, and the accuracy criterion, and thus allowing the selection of the irradiation parameters.

[0151] The node frequency chart can be stored in the memory of the adaptation module 14 or the simulation module 15, or in another memory of the non-invasive sensor 1.

[0152] Alternatively, the node frequency chart is generated from data obtained through in vivo experiments. In this embodiment, the simulation module 15 is not required.

[0153] In another embodiment, the node frequency chart is generated from both data obtained in silico using the simulation module and data obtained in vivo.

[0154] Process for measuring the parameter of interest A specific embodiment of a method for measuring a parameter of interest by the non-invasive sensor 1 based on indirect photoacoustic or photothermal detection is shown in FIG. 5. This includes the following steps.

[0155] a) Initialization: The adaptation module 14 selects an initial model configuration of the target layered environment 2 for irradiation based on a predetermined criterion. For example, in FIG. 5, the initial CMirrad model configuration is the CMk model configuration.

[0156] Various options are available for this initialization.

[0157] In particular, in the case of glucose measurements in interstitial fluid from a model configuration database and / or an experimental measurement database, an average set of level 1 and glucose parameters can be determined for a patient population or a given patient. This average parameter set corresponds to the initial model configuration CMk of the target layered environment 2, which has the highest probability of being most suitable for the next measurement in the absence of any other information, particularly in the absence of a measurement history.

[0158] b) Adaptation of irradiation parameters From the 1-node frequency chart, the processor of the adaptation module 14 determines a plurality of node modulation frequencies fnod associated with the group GMp of model configurations of the target layered environment to which the initial model configuration CMirrad belongs. 2 - Next, based on this plurality of node modulation frequencies, at least one specific modulation frequency foptim is determined. The at least one specific modulation frequency foptim can be the modulation frequency at which the sensor sensitivity is maximum in a modulation frequency range bounded by two (possibly consecutive) node modulation frequencies for the parameter of interest. In one embodiment, the at least one specific modulation frequency foptim is determined by the processor of the simulation module 15 or the adaptation module 14 based on the weighted average of at least two consecutive node modulation frequencies. 3 - And finally, based on these node modulation frequencies, a specific irradiation case for the model configuration CMirrad of the target layered environment 2 is determined, which is an irradiation case including at least a specific irradiation frequency foptim for the initial model configuration CMirrad. In particular, since the specific irradiation case is optimized with respect to the criteria for taking into account interference phenomena, it can be denoted as Iopt.k.

[0159] c) First irradiation: The target environment 2 is first irradiated, and at least one of the light sources 11a is configured according to a set of irradiation parameters corresponding to the specific irradiation case.

[0160] d) PA detection: The actual photoacoustic signal generated in response to the irradiation is detected by the photoacoustic detection cell 12.

[0161] If necessary, the signal processing module 13 receives and processes this actual photoacoustic signal, and after processing, transmits it to the adaptation module 14.

[0162] e) Determination of parameters of interest The processor of the adaptation module 14 determines the parameter of interest, in this example blood glucose, based on the detected photoacoustic or photothermal signal.

[0163] Several embodiments of step e) are possible.

[0164] In particular, the processor of the adaptation module 14 can implement a single inverse model.

[0165] Alternatively, the processor of the adapter module 14 can infer the blood glucose value from the detection signal using the law of motion.

[0166] In the most advanced embodiment shown in FIG. 5, step e) can be decomposed as follows.

[0167] e1) Solving the inverse problem The processor of the adaptation module 14 that implements the learned inverse model (in particular, at least the one corresponding to the model configuration CMk of the target layered environment 2 used for irradiation) receives, as input, the actual photoacoustic signal or photothermal signal and a specific irradiation case selected for irradiation, represented, for example, by Iopt.k, and determines the current target layered environment 2 model configuration and the parameter of interest, represented by CMmes (as shown in FIG. 7).

[0168] e2) Verification of the model configuration The adaptation module 14 compares the current layered target environment 2 model configuration CMmes with the model configuration CMirrad used for irradiation (model configuration CMk from the initialization step for the first measurement, which may be different for subsequent measurements, model configuration CM<).

[0169] Case 1) If the measured layered target environment 2 model configuration CMmes is identical to the model configuration CMirrad used for irradiation, the specific modulation frequency foptim and the resulting selected irradiation parameters, i.e., the multiple node modulation frequencies used to determine a specific irradiation case, are adapted to the measured physiological situation. In other words, the specific irradiation case selected is optimized to take into account the interference phenomena for the current physiological situation (which is not known in advance and keep in mind that it evolves over time), and the same applies to the inverse model used to determine the parameter of interest. As a result, the parameter of interest determined in step e1) by the adaptation module 14 is the measurement result.

[0170] Case 2) If the measured target layered environment model configuration 2 CMmes is different from the model configuration CMirrad, the adaptation module processor searches the node frequency chart for a new set of node modulation frequencies corresponding to the model configuration group GMp' to which the new target layered environment model configuration CMmes now belongs, and based on this set of node modulation frequencies, determines a specific new irradiation case including at least one modulation frequency foptim determined based on the set of node modulation frequencies for the target layered environment model configuration 2 CMmes. The specific new irradiation case is optimized to take into account the interference phenomena of the physiological situation, for example, represented as Iopt<. The adaptation module processor 14 transmits the corresponding parameters to the irradiation device 11.

[0171] Next, the target layered environment 2 is irradiated again according to the irradiation parameters of the irradiation case Iopt.< (subsequent irradiation). Next, the steps of detecting and, if necessary, processing the photoacoustic signal or the photothermal signal, step d), and solving the inverse problem, step e1), are repeated.

[0172] The process can also include step e2) for validating the model configuration.

[0173] Alternatively, the process can also include the repetition of step e2) for validating the model configuration.

[0174] In any case, in step e), the model configuration of the target layered environment 2, which is CMmes, may be different from the model configuration CMirrad. In fact, CMirrad is selected according to the initialization conditions (in this case, CMirrad is, in particular, the "average" model configuration in the case where there is no measurement history in the patient. This is the case of the first irradiation); or based on the previous measurement (when the measurement results following the previous irradiation are known and CMirrad is the model configuration most suitable for the patient).

[0175] Therefore, CMirrad is validated using the additional information obtained from the current irradiation, i.e., the signal detected by the photoacoustic detection cell 12. Incidentally, if CMmes = CMirrad, no further irradiation is required and the value of the parameter of interest is, in fact, the most accurate that can be obtained, but the model configuration validation step provides the confirmation that the additional information, i.e., the measurement accuracy, is actually maximum for this case.

[0176] In cases where the CMmes is different from the CMirrad, the adaptation of the irradiation parameters and the model configuration compensates for at least one additional irradiation, but still uses the energy consumption under control to increase the measurement accuracy by a second verification step and to enable confirmation that the measurement accuracy is actually at its maximum. When the verification step is allowed to be repeated (as shown in Figure 5), the process output obtains additional information that the measurement accuracy is actually at its maximum.

[0177] Generally, in this case, the measurement results of the parameter of interest are obtained after the first irradiation or after two irradiations. However, the number of steps required to verify the model configuration can be limited in order to ensure that the process converges and / or to limit the energy consumption.

[0178] Thus, in all cases, by selecting the irradiation case based on the node frequency chart, it is possible to limit the number of modulation frequencies and wavelengths used in the irradiation, provide non-redundant information for the current model configuration of the physiological target environment 2, retain only the values that are sufficient to obtain the desired measurement accuracy and / or optimal for limiting the sensor power consumption to a predetermined value.

[0179] In particular, one or more specific modulation frequencies of the source are selected to take into account the interference phenomenon, and the specific modulation frequency foptim is the modulation frequency at which the measurement sensitivity is the highest between two node modulation frequencies. By selecting an irradiation case with at least one specific modulation frequency, the sensitivity of the sensor is increased without increasing the power consumption of the sensor.

[0180] This can be seen from FIGS. 8a, 8b and 8c in which three consecutive situations were simulated to obtain three drawings. For all three drawings, the target layered environment 2 is the skin of a patient modeled by a two-layer model (stratum corneum, epidermis). The thickness of the upper layer (modeling the stratum corneum) is equal to 14 mm in the case of FIG. 8a, and when the measurements of FIGS. 8b and 8c are made, it is assumed that the event varies this thickness to develop and become equal to 18 mm. The thickness of the lower layer (modeling the epidermis) is considered infinite. The blood glucose is 0 in the upper layer. The upper layer contains 20% moisture and the lower layer contains 60% moisture.

[0181] The node frequency chart is used by the processor of the adaptation module 14 to determine a plurality of node frequencies for the initial situation (stratum corneum thickness equal to 14 mm) and to determine the irradiation case used for this situation based on these node frequencies. In this case, to obtain the measurement points shown in FIG. 8a, the skin model is irradiated (in a simulated manner) with a laser at a frequency of 1034 cm-1 modulated in intensity at three modulation frequencies foptim = 1580 Hz, fnod1 = 950 Hz and fnod2 = 4750 Hz determined from FIG. 3a.

[0182] Similarly, the node frequency chart enables the processor of the adaptation module 14 to determine a plurality of node modulation frequencies for the final situation (stratum corneum thickness equal to 18 mm) and to determine the irradiation case used for this situation based on these node frequencies. In this case, to obtain the measurement points shown in FIG. 8c, the skin model is irradiated (in a simulated manner) with a laser at a frequency of 1034 cm-1 modulated in intensity at three modulation frequencies foptim = 1000 Hz, f’nod1 = 600 Hz and f’nod2 = 3000 Hz determined from FIG. 3b.

[0183] In contrast, to obtain the measurement points shown in Fig. 8b, the skin model is irradiated with a laser of the intensity of the mantissa 1034 cm-1 modulated at three modulation frequencies specifically adapted to the skin model shown in Fig. 8a, namely foptim = 1580 Hz, fnod1 = 950 Hz, and fnod2 = 4750 Hz.

[0184] In the case where the irradiation case is selected based on the node modulation frequency adapted to the current physiological situation, that is, for Figs. 8a and 8c, it can be seen that the root mean square error (RMSE) between the predicted value and the actual value is acceptable (18.5 mg / dL and 15.4 mg / dL, respectively). On the other hand, when the node modulation frequency is inappropriate (as in Fig. 8b), the root mean square error (RMSE) between the predicted value and the actual value deteriorates significantly and reaches 28.4 mg / dL in this case.

[0185] Therefore, by selecting a limited number of modulation frequencies based on the node modulation frequency, it becomes possible to control the power consumption of the non-invasive sensor, but the adaptation of the node modulation frequency to the current physiological situation (based on which a specific modulation frequency is determined) also makes it possible to control the accuracy of the non-invasive sensor 1 so that, for example, the root mean square error or any other quantity capable of quantifying the accuracy is below a predetermined threshold.

[0186] The detection method can include, for example, the selection of each irradiation case of a single specific modulation frequency foptim determined based on a plurality of node frequencies by arithmetic mean or weighted mean. In particular, in the case of weighted mean, a larger weight can be assigned to the lowest node modulation frequency.

[0187] This configuration takes into account that the amplitude of the optothermal signal decreases with the modulation frequency of the incident light wave intensity.

[0188] As an example, in a specific case of photoacoustic measurement in the thermal piston regime described in Kottmann, J. et al. (2012); Glucose sensing in human epidermis using mid-infrared photoacoustic detection; Biomedical optics express, 3(4), 667-680, the amplitude of the photoacoustic signal is inversely proportional to the modulation frequency of the incident light wave. As a result, in particular in this case, a specific modulation frequency foptim can be taken as the centroid of two node modulation frequencies, each assigned a weight equal to the inverse of the other. Thus, when representing f1 and f2 as two consecutive node modulation frequencies, the specific modulation frequency can be calculated as follows: Formula 2 [Number]

[0189] In particular, in another embodiment suitable for the case of a closed photoacoustic detection cell, in order to take into account the frequency response of the detection cell, the specific modulation frequency foptim can be taken as the centroid of two consecutive node modulation frequencies, each assigned a weight equal to the inverse raised to the power of 3 / 2.

[0190] For example, additional weights can be selected to take into account the frequency response of one or more other elements in the measurement chain.

[0191] Alternatively, for each irradiation, the irradiation case selected in the process can include at least one node modulation frequency in addition to the specific foptim modulation frequency. In fact, the detected signal corresponding to the node modulation frequency does not carry information about the parameter of interest, but rather carries information about the structuring of the layered environment 2 independently of this parameter of interest. Thus, by selecting both the specific modulation frequency foptim and the node modulation frequency fnod, it can lead to the inverse problem being solved with higher efficiency and accuracy than any other selection of the two modulation frequencies.

[0192] In particular, as was done for the examples of FIGS. 8a, 8b and 8c, it is possible to select one or both of the node modulation frequencies that are closest to the determined optimal frequency foptim.

[0193] It is also possible to select two or more different specific modulation frequencies.

[0194] For example, if the main optimization criterion is energy consumption, i.e., if the non-invasive sensor is powered by a battery that reaches a predetermined lower charge threshold, it is possible to select only one modulation frequency for the irradiation, i.e., a specific foptim modulation frequency.

[0195] If the energy consumption criterion is not more important and irradiation at two different modulation frequencies is permitted, it is possible to select the specific modulation frequency foptim and the lowest of the node modulation frequencies used to determine foptim. In fact, it is this lowest node modulation frequency at which the signal amplitude is highest and thus the accuracy is best.

[0196] If the objective is to optimize the sensor sensitivity or accuracy with even lower energy consumption, the specific modulation frequency foptim and two node modulation frequencies can be used for the irradiation modulation frequencies to determine foptim and, optionally, second, third, etc. special modulation frequencies.

[0197] Finally, the present invention relates to a computer program comprising instructions for causing the non-invasive sensor 1 to execute the steps of a process according to any of the embodiments described above, according to one of the preceding embodiments.

Description of the reference numerals

[0198] 1 Non-invasive sensor based on indirect photoacoustic detection 11 Irradiation device 11a Light source 11b Intensity modulation device 11c Device for controlling the intensity modulation frequency fmod of the light source 11a 12 Detection cell 13 Signal processing module 14 Adaptation module 15 Simulation module 2A First layer of the target layered environment 2B Second layer of the target layered environment

Claims

1. A method for measuring a parameter of interest in a layered target environment (2) by a non-invasive sensor (1) based on photoacoustic detection or photothermal detection, comprising: a) providing a sensor, said sensor comprising: a light source (11a), a device for controlling a plurality of irradiation parameters of said light source (11a), said plurality of irradiation parameters including at least one frequency for modulating the intensity of said light source, a detection cell (12) configured to detect an acoustic signal or a thermal signal, a memory storing a node frequency chart, a plurality of model configuration groups (Gmp) of said layered target environment, each including at least two model configurations (Cmk) of said layered target environment in which only said parameter of interest differs from each other, and a plurality of irradiation case groups (GIq), each including at least two irradiation cases (Ij) that differ from each other only in the modulation frequency of the light source, and each irradiation case (Ij) including a set of irradiation parameter values, a plurality of node multiplets, each node multiplet including a characteristic common to all elements of a given model configuration group (Gmp) of said layered target environment and a plurality of associated node modulation frequencies, for said plurality of associated node modulation frequencies, the acoustic signal or thermal signal detected by said detection cell in response to irradiation by said light source exhibits a correlation with said parameter of interest that is less than a predetermined threshold, a processor, and an adaptation module (14) for exchanging information with said detection cell (12) and said light source irradiation parameter control device provided with a sensor, b) said adaptation module (14) selecting a model configuration (CMirrad) of the initial layered target environment for irradiation, c) the processor of said adaptation module (14) In the node frequency chart, determine a plurality of node modulation frequencies (f nod) associated with a model configuration group (G Mp) of the layered target environment to which the selected model configuration (C M irr ad) of the layered environment belongs. Next, based on the plurality of node modulation frequencies, determine at least one specific modulation frequency (f optim), where the specific modulation frequency is a weighted average of at least two of the plurality of node modulation frequencies and is bounded by two node modulation frequencies. Determine a specific irradiation case for the selected model configuration (C M irr ad) of the layered target environment, where the specific irradiation case includes at least one specific modulation frequency (f optim). d) irradiating the layered target environment (2) by the light source (11a) according to a set of the irradiation parameters of the specific irradiation case; e) detecting, by the detection cell (12), an acoustic signal or a thermal signal generated in response to the irradiation; f) determining, by the processor of the adaptation module (14), the parameter of interest based on the detected acoustic signal or thermal signal. A measuring method comprising the steps above. **Claim 2** The processor of the adaptation module (14) further implements an inverse modeling algorithm that receives, as inputs, an irradiation case (I j) and an acoustic signal or a thermal signal and provides, as outputs, a model configuration (C M k) of the layered target environment and a value of the parameter of interest. - Step f) comprises the following steps, namely: f1) The adaptation module processor receives, as inputs, the acoustic signal or the thermal signal detected by the detection cell (12) and the specific irradiation case used for the irradiation, and returns, as outputs, a model configuration (C M mes) of the current layered target environment and an estimated value (P est) of the parameter of interest from the inverse modeling algorithm. f2) The adaptation module processor (14) evaluates the selected layered target environment model configuration (CMirrad) by comparison with the current layered target model configuration (CMmes), and only if CMirrad is different from CMmes, g) The adaptation module (14) receives, as input, the model configuration (CMmes) of the current layered target environment and returns, as output, a new model configuration (CMirrad) of the selected layered target environment for irradiation, and then steps c), d), e) and f1) are repeated. f3) The value (Pmes) of the parameter of interest measured by the sensor is the last value of the estimated parameter of interest (Pest). The measuring method according to claim 1, comprising

3. The measuring method according to claim 1, further comprising that g) repeats f2) at the end of f1).

4. In advance, An I-processor, a database of model configurations including a multiplex (the model configuration (CMk) of the layered target environment, the irradiation case (Ij), the parameter of interest), and the acoustic or thermal signals detected by the detection cell (12) associated with each of the multiplexes, a node frequency chart is generated for a plurality of model configuration groups (GMp), and a plurality of irradiation case groups (GIq) and the node frequency chart includes a plurality of node multiplexes, each node multiplex including characteristics common to all elements of a model configuration group (GMp) of the layered target environment and a plurality of associated node modulation frequencies. This node frequency abacus is stored in the memory of the non-invasive sensor (1), the measuring method according to any one of claims 1 to 3.

5. ​The measurement method according to claim 4, comprising: the II-processor learns at least one inverse modeling algorithm from the model configuration database, and the at least one inverse modeling algorithm is stored in the memory of the non-invasive sensor (1).

6. The measurement method according to claim 4 or 5, wherein at least one of the acoustic signal or the thermal signal detected by the detection cell (12) associated with the multiplex stored in the model configuration database is simulated, that is, generated by a computerized simulation device.

7. The measurement method according to any one of claims 1 to 6, wherein the specific irradiation case for the selected model configuration determined by the adaptation module (14) in step c) further includes at least a node modulation frequency.

8. A non-invasive sensor (1) based on photoacoustic or photothermal detection, configured to measure a parameter of interest in a layered target environment (2), a light source (11a), a device for controlling the irradiation parameters of the light source (11a), a detection cell (12) configured to detect an acoustic signal or a thermal signal, a memory storing a node frequency abacus, a plurality of model configuration groups (Gmp) of the layered target environment, each including at least two model configurations (Cmk) of the layered target environment in which only the parameters of interest are different from each other, and A plurality of irradiation case groups (GIq), each including at least two irradiation cases (Ij) that differ from each other only in the modulation frequency of the light source, and each irradiation case (Ij) includes a set of irradiation parameter values. For the plurality of irradiation case groups (GIq), it includes a plurality of node multiplets, and each node multiplet includes a characteristic common to all elements of a given model configuration group (GMp) of the layered target environment and a plurality of related node modulation frequencies. For the plurality of related node modulation frequencies, the acoustic signal or thermal signal detected by the detection cell in response to irradiation by the light source exhibits a correlation with the parameter of interest that is less than a predetermined threshold value, a memory comprising The non-invasive sensor comprises a processor and further comprises an adaptation module (14) adapted to exchange information with the detection cell (12) and the light source irradiation parameter control device (11a), and the adaptation module (14) selecting a model configuration (CMirrad) of the initial layered target environment determining, in the node frequency chart, a plurality of node frequencies (f nod) associated with a model configuration group (GMp) to which the selected layered target environment model configuration (CMirrad) belongs determining at least one specific modulation frequency (f optim) based on the plurality of node modulation frequencies, wherein the specific modulation frequency is a weighted average of at least two of the plurality of node modulation frequencies and is bounded by two node modulation frequencies determining a specific irradiation case for the selected irradiation model configuration (CMirrad) including the at least one specific modulation frequency (f optim) transmitting the specific irradiation case to the light source irradiation parameter control device receiving an acoustic signal or a thermal signal detected by the detection cell (12) determining a value of the parameter of interest based on the detected acoustic signal or thermal signal and is further configured to do so. A non-invasive sensor.

9. The processor of the adaptation module: implements an inverse modeling algorithm that receives an irradiation case (Ij) including a set of irradiation parameters and an acoustic signal or a thermal signal as input, and provides a model configuration (CMk) of the layered target environment and a value of the parameter of interest as output; uses the inverse modeling algorithm to determine a current model configuration (CMmes) of the layered target environment and an estimated value (Pest) of the parameter of interest based on the received and detected photoacoustic signal or photothermal signal and the selected irradiation case; evaluates a selected model irradiation configuration (CMirrad) of the layered target environment by comparison with the current model configuration (CMmes) of the layered target environment; determines a new model irradiation configuration (CMirrad) of the selected layered target environment when receiving the current model irradiation configuration (CMmes) of the current layered target environment only if the selected model irradiation configuration (CMirrad) of the selected target layered environment is different from the current model configuration (CMmes) of the current target layered environment being compared; determines a plurality of node modulation frequencies (fnod) associated with a model configuration group (GMp) of the target layered environment to which the new model configuration (CMirrad) of the selected layered environment belongs in the node frequency chart; Next, based on the plurality of node modulation frequencies, determines at least one specific modulation frequency (foptim), the specific modulation frequency being a weighted average of at least two of the plurality of node modulation frequencies, bounded by two node modulation frequencies; determines a new specific irradiation case of the new model configuration (CMirrad) of the selected target layered environment, the specific irradiation case including the at least one specific modulation frequency (foptim); The light source (11a) irradiates the target layer environment (2) following the setting of the irradiation parameters for the new specific irradiation case, the detection cell (12) detects an acoustic signal or a thermal signal generated in response to this new irradiation, and the processor of the adaptation module (14) controls the irradiation parameters of the light source (11a) so as to determine the parameter of interest based on the newly detected acoustic signal or thermal signal. Transmitting the new specific irradiation case to the device for Determining the measured parameter of interest value (Pmes) based on the last estimated value of the parameter of interest (Pset). The non-invasive sensor according to claim 8, further adapted to do so.

10. The light source (11a), the device for controlling the irradiation parameters of the light source (11a), the detection cell (12) configured to detect an acoustic signal or a thermal signal, the memory storing the abacus of the node frequencies, and at least two of the devices of the adaptation module (14) are mechanically independent of each other. The non-invasive sensor according to claim 8 or 9.

11. A computer program comprising instructions for causing the non-invasive sensor (1) according to any one of claims 8 to 10 to execute the steps of the method according to any one of claims 1 to 7.

Images