Device for determining a vital sign of a user and method of determining a vital sign

Abstract

A device for determining a vital sign of a user (8) is specified, comprising: - a radiation emission unit (2) configured to emit radiation through the user's skin surface (80) during operation of the device (1); - a detector unit (3) sensitive to the radiation to be emitted by the radiation emission unit (2); - a temperature sensor (4); and - a controller unit (5) connected to the radiation emission unit (2), the detector unit (3), and the temperature sensor (4); wherein - the radiation emission unit (2) configured to emit radiation at a plurality of peak emission wavelengths (29A, 29B,..., 29N); and - the controller unit is configured to select one of the plurality of peak emission wavelengths (29A, 29B,..., 29N) based on a temperature signal obtained from the temperature sensor (4). Further, a method of determining a vital sign of a user (8) is specified.

Description

[0001] Description

[0002] DEVICE FOR DETERMINING A VITAL S IGN OF A USER AND METHOD OF DETERMINING A VITAL S IGN

[0003] The present application refers to a device for determining a vital sign of a user and to a method of determining a vital sign .

[0004] Photoplethysmographic ( PPG) signals can be used for non- invasively determining vital sign or health parameters . However, the signal quality of the PPG signal is often not suf ficient to extract valid features from the obtained data . In particular, reliability decreases i f the user' s tissue exhibits a comparably low perfusion .

[0005] An obj ect to be solved is to determine a vital sign in a more reliable manner .

[0006] This obj ect is inter alia solved by a device for determining a vital sign and a method according to the independent claims .

[0007] Further expediencies and configurations are subj ect of the dependent claims .

[0008] A device for determining a vital sign of a user is speci fied .

[0009] According to at least one embodiment , the device comprises a radiation emission unit configured to emit radiation through the user' s skin surface during operation of the device . Typically, the user is a human being . However, the user may also be an animal . For example , the radiation emission unit is configured to emit radiation onto a measuring spot on the user' s skin surface . For example , the measuring spot is located at the wrist , the fingertip, the forehead, an arm or a leg of the user .

[0010] For example , the radiation emission unit comprises a plurality of emitters . In particular, the emitters are controllable independently from one another . For example , at least one of the emitters comprises an active region based on a I I I-V compound semiconductor material .

[0011] I I I-V compound semiconductor materials are particularly suitable for radiation generation in the ultraviolet (AlxInyGai-x-y N) via the visible (AlxInyGai-X-yN, in particular for blue to green radiation, or AlxInyGai-X-yP, in particular for yellow to red radiation) to the infrared (AlxInyGai-x-yAs ) spectral range . Here , 0 < x < 1 , 0 < y < 1 and x + y < 1 apply in each case , in particular with x V 1 , y V I , x V 0 and / or y V 0 . With I I I-V compound semiconductor materials , in particular from the material systems mentioned, high internal quantum ef ficiencies can be achieved in radiation generation .

[0012] According to at least one embodiment of the device , the device comprises a detector unit sensitive to the radiation to be emitted by the radiation emission unit . For example , the detector unit comprises one or a plurality of detectors such as photodiodes . For example each of the detectors is in sensitive in at least part of the ultraviolet , visible and / or infrared spectral range . The detector unit may further comprise one or more filters to adapt the spectral and / or angular sensitivity of the detector unit or a detector thereof .

[0013] According to at least one embodiment of the device , the device comprises a temperature sensor . In particular, the temperature sensor is configured to be operable at least at temperatures of at least -30 ° C or at least -20 ° C or at least - 10 ° C and / or at most 40 ° C or at most 60 ° C .

[0014] According to at least one embodiment of the device , the device comprises a controller unit . In particular the controller unit is connected to the radiation emission unit and / or the detector unit and / or the temperature sensor . The connection to the controller unit may be direct or via one or more intervening electric or electronic components .

[0015] According to at least one embodiment of the device , the radiation emission unit is configured to emit radiation at a plurality of peak emission wavelengths . For example , two or more emitters of the radiation emission unit di f fer from one another with respect to their peak emission wavelengths .

[0016] According to at least one embodiment of the device , the controller unit is configured to select one of the plurality of peak emission wavelengths based on a temperature signal obtained from the temperature sensor . Thus , the radiation with the selected peak emission wavelength can be used as radiation source to irradiate the user' s skin and to detect the associated signal returning from the user' s tissue using the detector unit .

[0017] In at least one embodiment of the device , the device for determining a vital sign of a user comprises a radiation emission unit configured to emit radiation through the user' s skin surface during operation of the device and a detector unit sensitive to the radiation to be emitted by the radiation emission unit . The device further comprises a temperature sensor and a controller unit connected to the radiation emission unit , the detector unit and the temperature sensor . The radiation emission unit is configured to emit radiation at a plurality of peak emission wavelengths and the controller unit is configured to select one of the plurality of peak emission wavelengths based on a temperature signal obtained from the temperature sensor . For example , the selection is based on a machine learning algorithm using the temperature signal and a signal from the detector unit as input .

[0018] Thus , the radiation used for the determination of the vital sign is selected based on the temperature signal provided by the temperature sensor . It has been found that a temperature dependent selection of the peak emission wavelength results in a signi ficantly improved reliability of the signal used to determine the vital sign . It has turned out that the penetration depth of radiation into human tissue signi ficantly depends on the wavelength . Thus , a desired penetration depth for the determination of the vital sign can be set by appropriately choosing the peak emission wavelength .

[0019] Further, it has been found that the ideal penetration depth depends on the temperature . For example , at room temperature the perfusion directly below the epidermis is typically comparably good so that radiation having a small penetration depth can be used as a radiation source . However, the perfusion beneath the epidermis decreases with decreasing temperature . Consequently, radiation that penetrates deeper into the tissue reaches vessels that carry a higher volume of blood . This leads to an increased accuracy of the vital sign determination .

[0020] Further the signal-to-noise ratio is improved so that the measurements can be made at a lower sample rate and / or with lower radiation power so that the overall energy consumption may be reduced .

[0021] According to at least one embodiment of the device , the vital sign is derived from a PPG signal obtained using the detector unit . With the appropriate peak emission wavelength, PPG algorithms may be used to reliably obtain the vital sign to be determined .

[0022] According to at least one embodiment of the device , the selection of the peak emission wavelength is based on an algorithm using the temperature signal and a signal from the detector unit as input .

[0023] Depending on the signal quality and the outside conditions , the algorithm may cause the device to switch to a more suitable peak emission wavelength . The algorithm may be a conventional algorithm or a machine learning algorithm .

[0024] For example , the algorithm is a machine learning algorithm which, for example , analyses a shape of the measured PPG signal and identi fies changes in the PPG shape , for example due to the perfusion of di f ferent blood vessels . The algorithm may be further trained to user-speci fic characteristics such as the perfusion of the user' s tissue at the measuring spot used for the determination of the vital sign. Thus, user-specific characteristics and / or the position of the measuring spot chosen by the user can be taken into account for the determination of the vital sign, in particular automatically. This helps to further improve the reliability of the determination of the vital sign, even during operation of the device. In particular, the reliability can be increased compared to conventional devices where such user-specific characteristics are not addressed specifically.

[0025] According to at least one embodiment of the device, the vital sign is at least one of: heart rate, blood pressure, respiration rate, pulse rate variability. Further vital or health parameters may be derived from one or more of the above vital signs.

[0026] According to at least one embodiment of the device, at least two of the plurality of peak emission wavelengths differ from one another by at least 50 nm or at least 100 nm or at least 200 nm. Further peak emission wavelengths may be located between these two peak emission wavelengths having the above spectral distance. Using peak emission wavelengths with sufficiently large spectral distance may provide penetration depths significantly differing from one another.

[0027] According to at least one embodiment of the device, at least one of the plurality of peak emission wavelengths is in a range from 400 nm to 590 nm. For example, radiation in the green spectral range may provide a reliable signal at typical room temperatures (e.g. in a temperature range from 20°C to 25°C) . It has been found that radiation with a comparably low peak emission wavelength is more stable against motion arti facts , in particular compared to radiation having a larger wavelength and penetrating deeper into the tissue .

[0028] According to at least one embodiment of the device , at least one of the plurality of peak emission wavelengths is in a range from 600 nm to 1000 nm . In this spectral range the penetration depth typically varies from about 1 mm to about 3 mm . Typically, the highest penetration depths are obtained at wavelengths in a range from 780 nm to 940 nm . With further increasing wavelengths , the penetration depth decreases .

[0029] Radiation with a large penetration depth is particularly suitable at low temperatures where blood vessel constriction results in a decreased blood flow beneath the epidermis .

[0030] Thus , radiation penetrating deeper into the tissue reaches a higher blood volume , resulting in improved signal-to-noise ratio of the detector signal .

[0031] According to at least one embodiment of the device , the radiation emission unit comprises at least one tunable emitter, for example at least one tunable laser . For example , the peak emission wavelength of the tunable emitter can be tuned by at least 10 nm or at least 20 nm or at least 40 nm .

[0032] Using a tunable emitter, di f ferent peak emission wavelengths may be covered by a single emitter .

[0033] According to at least one embodiment of the device , the device is a wearable device . For example , the device is a smartwatch, a smart ring, a smart badge , an armlet , a patch, a headband, a helmet , a headset , earphones or glasses , for example augmented reality (AR) or virtual reality (VR) glasses . Further, a method of determining a vital sign of a user is speci fied . For example , the method can be performed using the device described above so that features described in connection with the device likewise apply to the method and vice versa .

[0034] According to at least one embodiment of the method, the method comprises the steps of obtaining a temperature signal associated with a measuring spot on the user' s skin surface irradiating the measuring spot with a radiation and detecting a portion of the radiation returning from the measuring spot wherein a peak emission wavelength of the radiation is selected based on the temperature signal .

[0035] Consequently, the radiation used for the determination of the vital sign can be adapted to the environment ' s temperature in the vicinity of the measuring spot . In particular, the peak emission wavelength may be changed during the method .

[0036] Compared to conventional methods where the same wavelength is always used for the determination of the speci fic vital sign, the reliability of the determination is increased .

[0037] According to at least one embodiment of the method, the peak emission wavelength is selected depending on its penetration depth into human tissue . Thus , the measuring volume can be adapted to temperature-dependent characteristics of the blood flow underneath the user' s skin surface .

[0038] According to at least one embodiment of the method, the peak emission wavelength is selected based on an algorithm using the temperature signal and a signal obtained from the returning radiation as input . For example , the algorithm is a machine learning algorithm analyzing a shape of the measured signal from the returning radiation . Alternatively, a conventional algorithm may be used .

[0039] For example , the method can be used for heart rate sensing, pulse rate reliability sensing, blood pressure sensing, oxygen saturation sensing ( SPO2 ) , body battery monitoring or respiratory rate monitoring .

[0040] Features described above in connection with at least one embodiment of the device or the method can be combined with other features described in connection with at least one embodiment of the device or the method unless they are contradictory .

[0041] Further configurations and expediencies will become apparent from the subsequent description of the exemplary embodiments in connection with the figures .

[0042] In the exemplary embodiments and figures , similar or similarly acting constituent parts are provided with the same reference signs . Generally only the di f ferences with respect to the exemplary embodiments are described . Unless otherwise speci fied, the description of a part or aspect in one exemplary embodiment applies to a corresponding part or aspect in another exemplary embodiment as well .

[0043] In the figures :

[0044] Figure 1A shows an exemplary embodiment of a device ;

[0045] Figures IB and 1C schematically show the interaction of the device with human tissue at warm temperatures ( Figure IB ) and at cold temperatures ( Figure 1C ) ; Figures 2A to 2C show an exemplary embodiment of a device wherein Figure 2A shows a perspective view onto the device , Figure 2B shows a view onto a rear side of the device , and Figure 2C shows an exemplary embodiment of electrical connections of the individual components of the device ; and

[0046] Figure 3 shows an exemplary embodiment of a method of determining a vital sign .

[0047] The figures are schematic representations . Thus , the elements illustrated in the figures and their si ze relationships among one another are not necessarily true to scale . Rather individual elements or layer thicknesses may be represented with an exaggerated si ze for the sake of better representability and / or for the sake of better understanding .

[0048] Figure 1A schematically illustrates a device 1 for determining a vital sign of a user 8 . The device 1 comprises a radiation emission unit 2 with a plurality of emitters 21A, 21B, ..., 21N . During operation of the device 1 , the radiation emitted by the radiation emission unit is emitted through a skin surface 80 into the user' s 8 tissue underneath the epidermis .

[0049] The device 1 further comprises a detector unit 3 sensitive to the radiation emitted by the radiation emission unit 2 during operation of the device 1 . In the exemplary embodiment shown, the detector unit 3 comprises a plurality of detectors 31A, 31B, ..., 31N wherein the number N of detectors 31 corresponds to the number of emitters 21 .

[0050] However, the number of detectors 31 may also be smaller than the number of emitters 21 . In particular, even a single detector 31 may be suf ficient to detect the radiation emitted by all emitters 21 of the emission unit 2 .

[0051] The device 1 further comprises a temperature sensor 4 and a controller unit 6 wherein the controller unit is connected to the radiation emission unit 2 , the detector unit 3 and the temperature sensor 4 .

[0052] The radiation emission unit 2 is configured to emit radiation at a plurality of peak emission wavelengths . In the exemplary embodiment shown, emitter 21A emits the radiation 91a having the shortest peak emission wavelength . For example , the peak emission wavelength of radiation 29A is at most 590 nm, for example in the blue spectral range or in the green spectral range .

[0053] The emitter 21N emits radiation with the largest peak emission wavelength 29N . For example , the peak emission wavelength is at least 600 nm, for example in a range from 600 nm to 1000 nm .

[0054] The controller unit 6 is configured to select one of the plurality of peak emission wavelengths based on a temperature signal obtained from the temperature sensor 4 during operation .

[0055] Thus , di f ferent peak emission wavelengths may be used for the determination of the speci fic vital sign to be obtained, depending on the temperature .

[0056] Figures IB and 1C illustrate the perfusion of human tissue beneath the epidermis 81 . At warm temperatures , for example at typical room temperature ( 20 ° C to 25 ° C ) , blood vessel dilation results in increased blood flow towards the blood vessels 82 located near the epidermis 81 . Thus , a comparably short penetration depth of radiation 29A is suf ficient to obtain a strong interaction with the radiation 29A with the blood within the blood vessels 82 so that the vital sign can be reliably derived using the PPG signal obtained from the detector unit 3 .

[0057] At colder temperatures , in contrast , the perfusion directly below the epidermis 81 decreases , as illustrated in Figure 1C . This is because a blood vessel constriction 83 as schematically illustrated results in decreased blood flow towards the skin surface 80 in order to reduce the heat loss across the epidermis 81 . In this case , radiation having a comparably small penetration depth interacts with a reduced blood volume so that the reliability of the PPG signal decreases .

[0058] By switching to radiation having a larger peak emission wavelength 29N and thus having a larger penetration depth, the interaction with blood can be increased, resulting in an improved signal-to-noise ratio .

[0059] The penetration depth can be set by an appropriate selection of the peak emission wavelength .

[0060] In the wavelength range from 150 to 380 nm (ultraviolet radiation) , the penetration depth is typically below 0 . 1 mm .

[0061] In the wavelength range from 390 nm to 470 nm (violet to deep blue radiation) , the penetration depth is about 0 . 3 mm . In the wavelength range from 475 to 545 nm (blue to green radiation) , the penetration depth is about 0 . 3 to 0 . 5 mm .

[0062] In the wavelength range from 545 nm to 600 nm ( yellow to orange radiation) , the penetration depth is about 0 . 5 mm to 1 . 0 mm .

[0063] In the wavelength range from 600 nm to 650 nm ( red radiation) , the penetration depth is about 1 . 0 to 2 . 0 mm

[0064] In a wavelength range from 650 nm to 950 nm ( deep red to near infrared radiation) , the penetration depth is about 2 mm to 3 mm .

[0065] In the wavelength range from 950 nm to 1200 nm (near infrared radiation) the penetration depth is about 1 mm .

[0066] For example , light in the green spectral range has been proven to be particularly suitable for PPG measurements on the wrist of the user 8 as it only superficially penetrates into the skin and is less prone to motion artefacts .

[0067] At colder temperatures , however, more reliable results are obtained at larger peak emission wavelengths having larger penetration depths . For example , colder temperatures may occur during outdoor activities such as skiing or mountaineering . In particular, radiation in the wavelength range from 650 nm to 1000 nm is particularly suited for measurements at lower temperatures .

[0068] Consequently, by selecting an appropriate peak emission wavelength from a plurality of available peak emission wavelengths ( 29A, 29B, ... , 29N) , the determination of the vital sign to be monitored can be reliably obtained over a large temperature range .

[0069] For example , the vital sign to be determined is the heart rate , the blood pressure , the respiration rate or the pulse rate variability . Further vital or health parameters may be derived from these vital signs .

[0070] The selection of the appropriate peak emission wavelength based on the temperature signal can be based on an algorithm using the temperature signal and a signal from the detector unit as input . The algorithm may be a conventional algorithm or a machine learning algorithm .

[0071] With the help of arti ficial intelligence , the shape of the PPG signal can be analyzed and changes in the PPG shape due to a small change of the perfusion of di f ferent blood vessels can be identi fied . Consequently, a highly appropriate selection of the suitable peak emission wavelengths can be obtained . Furthermore , the selection may also be adapted to the user-speci fic perfusion or the location-speci fic perfusion at the measuring spot .

[0072] A distance between the temperature sensor 4 and the radiation emission unit 2 is preferably at most 5 cm . A comparably small distance ensures that there is suf ficient correlation between the measured temperature signal and the perfusion at the measuring spot for the selection of the most appropriate peak emission wavelength ( 29A, 29B, ..., 29N) . For example , the distance between the temperature sensor and the measuring spot is at most 5 cm . The number N of peak emission wavelengths ( 29A, 29B, 29N) does not necessarily correspond to the number of individual emitters 21 of the radiation emission unit 2 . For example , the radiation emission unit 2 may comprise at least one tunable emitter 21 , for example a tunable laser that covers two or more di f ferent peak emission wavelengths ( 29A, 29B, ..., 29N) .

[0073] The number N of peak emission wavelengths may, for example , be at least 2 or at least 3 or at least 4 and / or at most 20 or at most 10 , wherein the peak emission wavelengths ( 29A, 29B, ..., 29N) may mutually di f fer from one another by at least 5 nm .

[0074] At least two of the plurality of peak emission wavelengths ( 29A, 29B, ..., 29N) may di f fer from one another by at least 50 nm or at least 100 nm or at least 200 nm .

[0075] The device 1 is preferably a wearable device , for example a smartwatch, a smart ring, a smart badge , an armlet , a headband, a helmet , a headset , glasses , a patch or earphones .

[0076] Figures 2A to 2C illustrate a further exemplary embodiment of a device 1 wherein the exemplary embodiment substantially corresponds to the exemplary embodiment described in connection with Figures 1A to 1C .

[0077] As illustrated in Figure 2A, the device 1 is a smartwatch so that a measuring spot 19 is located at or near the user' s 8 wrist .

[0078] As illustrated in Figure 2B showing the rear side 10 of the device 1 which is in contact with the user' s skin or which at least faces the user' s skin during operation, the device 1 comprises a radiation emission unit 2 with four emitters 21 emitting radiation at di f ferent peak emission wavelengths ( 29A, 29B, 29N) .

[0079] For example , at least one emitter 21 emits in the green spectral range and / or at least one emitter 21 emits in the red spectral range and / or at least one emitter 21 emits in the near infrared spectral range , for example in a wavelength range from 700 nm to 1000 nm .

[0080] Further, the device 1 comprises a detector unit 3 which is configured to be sensitive to the radiation emitted by the four emitters 21 of the radiation emission unit 2 . As described in connection with Figures 1A to 1C, the number of emitters 21 and the number of detectors 31 within the detector unit 3 may be varied in wide limits .

[0081] Figure 2C schematically illustrates the electrical connections between the individual components of the device 1 . The electrical connection indicated by arrows may be direct or via further intervening elements such as electric or electronic components . This scheme may also be used for other wearable devices , such as those described in connection with Figures 1A to 1C .

[0082] The device 1 comprises an optical unit 7 including the radiation emission unit 2 and the detector unit 3 . A controller unit 5 is connected to the optical unit 7 via an optional analog front end 6 .

[0083] The device 1 further comprises a temperature sensor 4 and a memory 75 connected to the controller unit 5 . As described in the previous exemplary embodiment , the controller unit 5 is configured to select one of the peak emission wavelengths provided by the emitters 21 based on the temperature signal provided by the temperature sensor 4 .

[0084] Figure 3 illustrates a method of determining a vital sign of a user according to an exemplary embodiment .

[0085] In a step 901 a temperature signal associated with a measuring spot on the user' s skin surface is obtained . For example , a distance between a temperature sensor and the measuring spot is at most 5 cm .

[0086] In a step 902 , the measuring spot is irradiated with radiation . In a step 903 , a portion of the radiation returning from the measuring spot is detected . A peak emission wavelength of the radiation used in step 902 is selected based on the temperature signal obtained in step 901 . For example , the method is performed using a device as described in connection with Figures 1A to 1C or 2A to 2C .

[0087] In particular, the peak emission wavelength can be selected depending on its penetration depth into human tissue .

[0088] Further, the selection of the peak emission wavelength can be based on a conventional or machine learning algorithm using the temperature signal and a signal detecting the portion of the radiation returning from the measuring spot 19 as input .

[0089] As described in connection with the previous exemplary embodiments , the algorithm may be a machine learning algorithm analyzing a shape of the measured signal from the detector unit . As described in connection with the device , the method allows to reliably obtain a vital sign to be determined . In particular, a high reliability is obtained over a large temperature range so that the vital sign can also be obtained at cold temperatures , for example during outdoor activities .

[0090] This patent application claims the priority of German patent application 10 2023 115 452 . 4 , the disclosure content of which is hereby incorporated by reference .

[0091] The invention described herein is not restricted by the description given with reference to the exemplary embodiments . Rather, the invention encompasses any novel feature and any combination of features , including in particular any combination of features in the claims , even i f this feature or this combination is not itsel f explicitly indicated in the claims or exemplary embodiments .

[0092] References

[0093] 1 device

[0094] 10 rear side

[0095] 19 measuring spot

[0096] 2 radiation emission unit

[0097] 21 , 21A, 21B, 21N emitter

[0098] 29A, 29B, 29N radiation with peak emission wavelength

[0099] 3 detector unit

[0100] 31 detector

[0101] 4 temperature sensor

[0102] 5 controller unit

[0103] 6 analog front end

[0104] 7 optical unit

[0105] 75 memory

[0106] 8 user

[0107] 80 skin surface

[0108] 81 epidermis

[0109] 82 blood vessel

[0110] 83 constriction

[0111] 901 , 902 , 903 step

Claims

Claims1. A device for determining a vital sign of a user (8) , comprising :- a radiation emission unit (2) configured to emit radiation through the user's skin surface (80) during operation of the device ( 1 ) ;- a detector unit (3) sensitive to the radiation to be emitted by the radiation emission unit (2) ;- a temperature sensor (4) ; and- a controller unit (5) connected to the radiation emission unit (2) , the detector unit (3) , and the temperature sensor (4) ; wherein- the radiation emission unit (2) configured to emit radiation at a plurality of peak emission wavelengths (29A,29B, ... , 29N) ;- the controller unit is configured to select one of the plurality of peak emission wavelengths (29A, 29B, ..., 29N) based on a temperature signal obtained from the temperature sensor ( 4 ) ; and- the selection is based on a machine learning algorithm using the temperature signal and a signal from the detector unit (3) as input.

2. The device according to claim 1, wherein the vital sign is derived from a photoplethysmographic (PPG) signal obtained using the detector unit (3) .

3. The device according to claim 2, wherein the machine learning algorithm is configured to analyse a shape of thePPG signal and to identify changes in the shape of the PPG signal .

4. The device according to claim 3, wherein the machine learning algorithm is configured to be trained to user-specific characteristics of a perfusion of the user' s tissue at a measuring spot used for the determination of the vital sign.

5. The device according to any of the preceding claims, wherein the vital sign is at least one of: heart rate, blood pressure, respiration rate, pulse rate variability.

6. The device according to any of the preceding claims, wherein at least two of the plurality of peak emission wavelengths (29A, 29B, ... , 29N) differ from one another by at least 50 nm .

7. The device according to any of the preceding claims, wherein at least one of the plurality of peak emission wavelengths (29A, 29B, ... , 29N) is in a range from 400 nm to 590 nm .

8. The device according to any of the preceding claims, wherein at least one of the plurality of peak emission wavelengths (29A, 29B, ... , 29N) is in a range from 600 nm to 1000 nm.

9. The device according to any of the preceding claims, wherein the radiation emission unit (2) comprises at least one tunable emitter (21) .

10. The device according to any of the preceding claims, wherein the device is a wearable device.

11. A method of determining a vital sign of a user (8) , comprising the steps of: a) obtaining a temperature signal associated with a measuring spot (19) on the user's skin surface (80) ; b) irradiating the measuring spot (19) with a radiation; and c) detecting a portion of the radiation returning from the measuring spot; wherein a peak emission wavelength (29A, 29B, ... , 29N) of the radiation is selected based on a machine learning algorithm using the temperature signal and a signal obtained from the returning radiation as input.

12. The method according to claim 11, wherein the peak emission wavelength (29A, 29B, ... , 29N) is selected depending on its penetration depth into human tissue .

13. The method according to claim 11 or 12, wherein the machine learning algorithm is trained to userspecific characteristics of the perfusion of the user's tissue at a measuring spot used for the determination of the vital sign.

14. The method according to claim 13, wherein the machine learning algorithm analyzes a shape of the signal from the returning radiation.

15. The method according to any one of claims 11 to 14, wherein the method is performed using a device according to any one of claims 1 to 10.

Images