Physiological parameter detection method and device, wearable equipment and storage medium

Abstract

The invention relates to a physiological parameter detection method and device, wearable equipment and a storage medium, and the physiological parameter detection method is applied to the wearable equipment and comprises the steps that a physiological state parameter set of a user is acquired, and the physiological state parameter set comprises physiological state parameters of multiple dimensions, the physiological state parameters of the multiple dimensions are used for determining to-be-detected target physiological parameter values, and the target physiological parameter values comprise physiological parameter values of the user in a motion state and / or a non-motion state; a target physiological parameter value is determined, the target physiological parameter value is determined according to the physiological state parameters subjected to normalization processing and a target physiological parameter detection model, and the physiological state parameters subjected to normalization processing are obtained through normalization processing on the basis of the multiple physiological state parameters in the physiological state parameter set. Through the physiological parameter detection method and device, the physiological parameter detection method has higher robustness and universality.

Description

Technical Field

[0001] The present disclosure relates to the technical field of wearable devices, and in particular to a physiological parameter detection method, device, wearable device and storage medium. Background Art

[0002] Physiological parameters are important indicators of whether the human body's physiological functions are healthy. For example, blood oxygen saturation is an extremely important diagnostic indicator. For people who exercise regularly, high-intensity exercise will increase the body's oxygen consumption in a short period of time. If the blood oxygen saturation in the human body can be monitored in time and timely reminders can be given, the danger can be greatly reduced. Therefore, in related technologies, it is necessary to detect human physiological parameters.

[0003] In the related technologies, physiological parameters are mainly detected in a static state through invasive detection methods and non-invasive detection methods. Among them, the invasive detection method is complicated to operate and has high detection costs. Although the non-invasive operation is simple, both cannot perform continuous and accurate detection in a moving state and are not easy to carry. Summary of the invention

[0004] In order to overcome the problems existing in the related art, the present disclosure provides a physiological parameter detection method, apparatus, wearable device and storage medium.

[0005] According to a first aspect of an embodiment of the present disclosure, a physiological parameter detection method is provided, the method comprising obtaining a physiological state parameter set of a user, the physiological state parameter set comprising physiological state parameters of multiple dimensions, the physiological state parameters of multiple dimensions being used to determine a target physiological parameter value to be detected, the target physiological parameter value comprising a physiological parameter value of the user in a motion state and / or a non-motion state; normalizing the multiple physiological state parameters in the physiological state parameter set to obtain normalized physiological state parameters; determining a target physiological parameter value, the target physiological parameter value being determined based on the normalized physiological state parameters and a target physiological parameter detection model, the normalized physiological parameter being obtained by normalizing the multiple physiological state parameters in the physiological state parameter set.

[0006] In one embodiment, the target physiological parameter value is determined according to the normalized physiological state parameter and the target physiological parameter detection model in the following manner: Kalman filtering is performed on the target physiological parameter detection value corresponding to the output result to obtain the target physiological parameter value; the output result is the output result corresponding to the normalized physiological state parameter of the target physiological parameter detection model as the input.

[0007] In one embodiment, the target physiological parameter detection value corresponding to the output result is subjected to Kalman filtering to obtain the target physiological parameter value, including: determining the Kalman filter gain of the kth Kalman filtering of the target physiological parameter detection value based on the estimation error and prediction error of the k-1th Kalman filtering of the target physiological parameter detection value; updating the prediction error based on the Kalman filter gain and the estimation error of the k-1th Kalman filtering, and determining the kth target physiological parameter estimation value based on the Kalman filter gain, the target physiological parameter detection value and the target physiological parameter estimation value of the k-1th Kalman filtering; and using the kth target physiological parameter estimation value as the target physiological parameter value.

[0008] In one embodiment, the target physiological parameter detection model is determined in the following manner: a first training sample is obtained, wherein the first training sample includes physiological state sample parameters of multiple dimensions included in the physiological state parameter set; the physiological state sample parameters of multiple dimensions are normalized to obtain a second training sample; based on the second training sample, the target physiological parameter detection model is constructed using a radial basis function as a kernel function.

[0009] In one embodiment, the target physiological parameter detection model is constructed based on the second training sample and with the radial basis function as the kernel function, including: training the target physiological parameter detection prediction model based on N groups of the second training samples to obtain N cross-validation root mean square errors, the N cross-validation root mean square errors corresponding to N groups of penalty factors and kernel function parameters, the N groups of penalty factors and kernel function parameters corresponding to N target physiological parameter detection prediction models; based on grid optimization, determining the minimum cross-validation root mean square error among the N cross-validation root mean square errors, and determining a group of penalty factors and kernel function parameters corresponding to the minimum cross-validation root mean square error as the target penalty factor and target kernel function parameters; selecting the target physiological parameter detection prediction model corresponding to the target penalty factor and the target kernel function parameters as the target physiological parameter detection model.

[0010] According to a second aspect of an embodiment of the present disclosure, a physiological parameter detection device is provided, the device comprising: an acquisition unit, used to acquire a physiological state parameter set of a user, the physiological state parameter set comprising physiological state parameters of multiple dimensions, the physiological state parameters of multiple dimensions being used to determine a target physiological parameter value to be detected, the target physiological parameter value comprising the physiological parameter value of the user in a motion state and / or a non-motion state; a processing unit, used to determine the target physiological parameter value, the target physiological parameter value being determined based on normalized physiological state parameters and a target physiological parameter detection model, the normalized physiological state parameters being obtained by normalizing the multiple physiological state parameters in the physiological state parameter set.

[0011] In one embodiment, the processing unit determines the target physiological parameter value based on the normalized physiological state parameters and the target physiological parameter detection model in the following manner: Kalman filtering is performed on the target physiological parameter detection value corresponding to the output result to obtain the target physiological parameter value; the output result is the output result corresponding to the normalized physiological state parameter of the target physiological parameter detection model whose input is the output result.

[0012] In one embodiment, the processing unit performs Kalman filtering on the target physiological parameter detection value corresponding to the output result to obtain the target physiological parameter value in the following manner: determining the Kalman filter gain of the kth Kalman filtering of the target physiological parameter detection value based on the estimation error and prediction error of the k-1th Kalman filtering of the target physiological parameter detection value; updating the prediction error based on the Kalman filter gain and the estimation error of the k-1th Kalman filtering, and determining the kth target physiological parameter estimation value based on the Kalman filter gain, the target physiological parameter detection value and the target physiological parameter estimation value of the k-1th Kalman filtering; and using the kth target physiological parameter estimation value as the target physiological parameter value.

[0013] In one embodiment, the processing unit is also used to determine the target physiological parameter detection model; the processing unit determines the target physiological parameter detection model in the following manner: obtaining a first training sample, the first training sample includes physiological state sample parameters of multiple dimensions included in the physiological state parameter set; normalizing the physiological state sample parameters of multiple dimensions to obtain a second training sample; based on the second training sample, using the radial basis function as the kernel function, constructing the target physiological parameter detection model.

[0014] In one embodiment, the processing unit is also configured to construct the target physiological parameter detection model based on the second training sample and the radial basis function as the kernel function: train the target physiological parameter detection prediction model based on N groups of the second training samples to obtain N cross-validation root mean square errors, the N cross-validation root mean square errors correspond to N groups of penalty factors and kernel function parameters, and the N groups of penalty factors and kernel function parameters correspond to N target physiological parameter detection prediction models; based on grid optimization, determine the minimum cross-validation root mean square error among the N cross-validation root mean square errors, and determine a set of penalty factors and kernel function parameters corresponding to the minimum cross-validation root mean square error as the target penalty factor and target kernel function parameters; select the target physiological parameter detection prediction model corresponding to the target penalty factor and the target kernel function parameters as the target physiological parameter detection model.

[0015] According to a third aspect of an embodiment of the present disclosure, a wearable device for physiological parameter detection is provided, comprising: a sensor group for acquiring physiological state parameters of multiple dimensions, wherein the physiological state parameters of multiple dimensions are used to determine target physiological parameter values ​​to be detected, wherein the target physiological parameter values ​​include physiological parameter values ​​of a user in a motion state and / or a non-motion state; a processor; and a memory for storing processor executable instructions; wherein the processor is configured to: determine a target physiological parameter value that matches the physiological state parameter, wherein the target physiological parameter value is obtained by executing the method described in the first aspect or any one of the embodiments of the first aspect.

[0016] In one embodiment, the sensor group is communicatively connected to the processor to send the physiological state parameters of multiple dimensions acquired by the sensor group to the processor; the processor determines the target physiological parameter value based on the physiological state parameters.

[0017] In one embodiment, the wearable device also includes: a communication component; the communication component is used to send the physiological state parameter to an electronic device different from the wearable device, and receive a target physiological parameter value determined by the electronic device based on the physiological state parameter, and send the target physiological parameter value to the processor.

[0018] In one embodiment, the sensor group includes one or more of the following sensors: a short-wave infrared sensor; a heart rate sensor; a temperature sensor; and an acceleration sensor.

[0019] In one embodiment, the wearable device is: headphones.

[0020] According to a fourth aspect of an embodiment of the present disclosure, a storage medium is provided, in which instructions are stored. When the instructions in the storage medium are executed by a processor of a terminal, the terminal is enabled to execute the method described in the first aspect or any one of the embodiments of the first aspect.

[0021] The technical solution provided by the embodiments of the present disclosure may include the following beneficial effects: According to the embodiments of the present disclosure, a set of physiological state parameters of the user in different states is obtained through a wearable device, wherein the set of physiological state parameters includes physiological state parameters of multiple different dimensions. The target physiological parameter value is determined, and the target physiological parameter value is determined according to the normalized physiological state parameter and the target physiological parameter detection model, wherein the normalized physiological state parameter is obtained by normalizing multiple physiological state parameters in the set of physiological state parameters. Furthermore, this solution can solve the problem of not being able to measure the physiological parameter value in the motion state, and the problem of inconvenience in carrying when measuring physiological parameters by instruments.

[0022] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0024] Figure 1a The present invention is an operation flow chart showing a method of testing physiological parameters of a human body by using a blood oximeter according to an exemplary embodiment.

[0025] Figure 1b The present invention is an operation flow chart showing a method of testing physiological parameters of a human body by using a blood oximeter according to an exemplary embodiment.

[0026] Figure 1c The present invention is an operation flow chart showing a method of testing physiological parameters of a human body by using a blood oximeter according to an exemplary embodiment.

[0027] Figure 2 is a flow chart of a physiological parameter detection method according to an exemplary embodiment.

[0028] Figure 3 The present invention is a flow chart showing a method of obtaining a target physiological parameter value through Kalman filtering according to an exemplary embodiment.

[0029] Figure 4 The present invention is a flow chart showing a method of obtaining a target physiological parameter value by using a Kalman filter algorithm to obtain a target physiological parameter detection value according to an exemplary embodiment.

[0030] Figure 5 The figure is a flow chart of a training target physiological parameter detection model according to an exemplary embodiment.

[0031] Figure 6 The present invention is a flowchart showing a method of constructing a target physiological parameter detection model based on a second training sample and taking a radial basis function as a kernel function according to an exemplary embodiment.

[0032] Figure 7 The present invention is a flowchart showing a method of determining a target penalty factor and target kernel function parameters in a grid optimization manner according to an exemplary embodiment.

[0033] Figure 8 The present invention is a scene diagram showing detection of a physiological parameter of a user based on a wearable device worn by the user according to an exemplary embodiment.

[0034] Fig. 9 It is a block diagram of a device for detecting physiological parameters according to an exemplary embodiment.

[0035] Fig.10 It is a block diagram of a device for physiological parameter detection according to an exemplary embodiment. DETAILED DESCRIPTION

[0036] Here, exemplary embodiments will be described in detail, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present disclosure.

[0037] The physiological parameter detection method provided by the embodiments of the present disclosure is applied to the scene of portable detection equipment, for example, in the scene of wireless Bluetooth headsets, etc., when a user wears Bluetooth headsets to run, skip rope, walk briskly, etc., physiological parameters of the human body, such as blood oxygen saturation, can be detected.

[0038] The current method of physiological parameter detection is to detect the physiological parameters of the human body in a calm state through invasive detection methods or non-invasive detection methods. However, the invasive test method has problems such as difficult simple operation, high detection cost, and inability to detect continuously. Although the non-invasive detection method is simple to operate, it is also unable to detect continuously. Neither method can timely and accurately test the physiological parameters of the human body in a state of motion. Alternatively, a blood oximeter can be used to test the physiological parameters in a state of motion, but the person may remain in a calm state during the test. For example, Figure 1a , Figure 1b as well as Figure 1cThe following are the operation flow charts of using a blood oximeter to test human physiological parameters. They include: resting for 5 minutes before measuring, rubbing the palms of both hands until they are warm. Figure 1a Place your left index finger into the oximeter. Figure 1b Keep Figure 1b The posture remains unchanged for 8 seconds and the data is checked. Figure 1c shown.

[0039] Furthermore, in the current physiological parameter detection, the following formula is used to calculate the physiological parameter value: SpO2 = a × (PI 660 / PI 940 ) 2 +b×(PI 660 / PI 940 )+c. However, the coefficients involved in this formula are all obtained by calibrating the sensor in a laboratory environment. Therefore, during use, you need to remain calm and detect the physiological parameter values ​​in a non-exercise state.

[0040] In view of this, the present disclosure provides a physiological parameter detection method, which obtains a set of physiological state parameters of the user in different states through a device. And determines the target physiological parameter value based on the normalized physiological state parameters and the target physiological parameter detection model. In this way, there is no need to wait for the user to be in a calm state to measure the physiological parameter value, so it can be applied to the user in a state of motion or in a calm state, and has higher universality. Among them, a physiological parameter detection method provided by the present disclosure can be a blood oxygen saturation detection method in some embodiments.

[0041] Figure 2 is a flow chart of a physiological parameter detection method according to an exemplary embodiment. Figure 2 As shown, the physiological parameter detection method is used in a wearable device, including the following steps.

[0042] In step S11, a set of physiological state parameters of the user is obtained.

[0043] In the disclosed embodiment, the physiological state parameter set includes multiple physiological state parameters of different dimensions, wherein the multiple physiological parameters of different dimensions are used to determine the target physiological parameter value to be detected.

[0044] In the disclosed embodiment, the target physiological parameter value is the physiological parameter value of the user when the user is in motion or in a non-motion state, wherein the physiological parameter value in a calm state can be determined based on the pulse perfusion index.

[0045] In an example of the present disclosure, multiple physiological state parameters of different dimensions in a physiological state parameter set of a user can be obtained through a wearable device worn by the user.

[0046] In the disclosed embodiment, the physiological parameter may be human blood oxygen saturation, which is a key indicator of whether human physiological functions are healthy and can measure the ability of human blood to carry oxygen.

[0047] In step S12, a target physiological parameter value is determined.

[0048] The target physiological parameter value is determined based on the normalized physiological state parameter and the target physiological parameter detection model. In addition, the normalized physiological state parameter can be obtained by normalizing multiple physiological state parameters in the physiological state parameter set.

[0049] In the embodiment of the present disclosure, since the physiological state parameter set includes multiple different physiological state parameters, and the multiple different physiological parameters have different dimensions and have large differences, the multiple different physiological state parameters in the state parameter set are normalized.

[0050] In an example of the present disclosure, multiple different physiological state parameters in the physiological state parameter set are normalized to obtain normalized physiological state parameters, which can convert physiological state parameters of different dimensions to the same value range and eliminate the correlation between physiological state parameters of different dimensions.

[0051] In the embodiment of the present disclosure, the target physiological parameter value is determined based on the normalized physiological state parameter and the target physiological parameter detection model. It can be understood that the acquired physiological state parameter set is normalized and the obtained physiological state parameter is input into the target physiological parameter detection model, wherein the target physiological parameter detection model is a model that has been trained in advance and embedded in the wearable device.

[0052] In the embodiment of the present disclosure, a physiological state parameter set of the current user can be acquired through a sensor, and the physiological state parameter set of the current user has been embedded in a target physiological parameter detection model in a wearable device, so that a target physiological parameter value can be obtained.

[0053] The physiological parameter detection method provided by the embodiment of the present disclosure obtains the physiological state parameter set of the user, and inputs the obtained physiological state parameter set into a predetermined target physiological parameter detection model to obtain the target physiological parameter value. In this way, it is not necessary to remain calm during the detection process, and it is easy to carry, and can monitor the physiological parameter value of the human body in real time in a moving state or a non-moving state.

[0054] Embodiments of the Present Disclosure The following describes a method for detecting physiological parameters according to an embodiment of the present disclosure.

[0055] The target physiological parameter may be the user's current blood oxygen saturation value, for example, the user's blood oxygen saturation value during exercise.

[0056] In the disclosed embodiment, the target physiological parameter value is obtained according to a predetermined target physiological parameter detection model. The physiological state parameter set obtained based on the wearable device is input into the target physiological parameter detection model, and the target physiological parameter value is output through the target physiological parameter model.

[0057] In one implementation of the disclosed embodiment, a physiological state parameter set is obtained, wherein the physiological state parameter set includes multiple physiological state parameters of different dimensions. For example, based on sensors in wearable devices, such as wireless Bluetooth headsets, a physiological state parameter set of a human body in motion is obtained. The physiological state parameter set includes: human heart rate, human body temperature, movement speed, blood oxygen saturation value in a calm state, and pulse perfusion index.

[0058] In one implementation, human heart rate, movement speed, human body temperature, and DC and AC components of a specific wavelength spectrum are obtained through sensors in a wearable device.

[0059] In one embodiment, the pulse perfusion index is based on the skin diffuse reflectance spectrum signal and is obtained using the following formula:

[0060] PI λ =λ AC / λ DC

[0061] In the above formula, PI λ is the pulse perfusion index, λ AC is the AC component of the diffuse reflectance spectrum signal with a wavelength of λ, DC is the DC component of the diffuse reflectance spectrum signal.

[0062] In one implementation of the disclosed embodiment, the physiological parameter value of the user in a non-exercise state included in the target physiological parameter value is the blood oxygen saturation value of the human body in a calm state, which is obtained based on the pulse perfusion index using the following formula:

[0063] SpO2=a×(PI 660 / PI 940 ) 2 +b×(PI 660 / PI 940 )+c

[0064] In the above formula, a, b, and c are coefficients of the formula, which are determined by the sensors in the wearable device. SpO2 is the first blood oxygen saturation value, PI 660 is the pulse perfusion index corresponding to the spectral signal at 660 nm, PI940 is the pulse perfusion index corresponding to the 940-nanometer spectral signal.

[0065] In the embodiment of the present disclosure, the following formula can be used to normalize multiple different physiological parameters in the physiological parameter set:

[0066] Z=(Z max -Z min )×(xx min ) / (x max -x min )+Z min

[0067] In the above formula, Z is the normalized variable, Z max is the upper limit of the normalized variable, Z min is the lower limit of the variable after normalization, x is the variable before normalization, and x max is the maximum value of the variable before normalization, x min The normalized physiological parameter is input into the target physiological parameter detection model, and the target physiological parameter value is output through the target physiological parameter model. For example, the normalized blood oxygen saturation parameter can be input into the blood oxygen saturation detection model, and the blood oxygen saturation detection model outputs the current blood oxygen saturation value of the user.

[0068] In the disclosed embodiment, the target physiological parameter value is determined based on the output result of the target physiological parameter detection model. The output result of the target physiological parameter detection model can be iterated according to Kalman filtering to obtain the target physiological parameter value.

[0069] In one example, a target physiological parameter detection value corresponding to the output result is subjected to Kalman filtering to obtain the target physiological parameter value. It can be understood that the second blood oxygen saturation obtained by the target physiological parameter detection model is further optimized by the Kalman filtering algorithm to obtain the target physiological parameter value.

[0070] In one example, the Kalman filter algorithm is an algorithm that uses the linear system state equation and the system input and output observation data to optimally estimate the system state. In particular, it can estimate and predict the state of dynamic behavior in real time. Figure 3 FIG. 1 is a flow chart showing a method of obtaining a target physiological parameter value through Kalman filtering according to an exemplary embodiment. Figure 3 As shown, the following steps are included.

[0071] In step S21, a Kalman filter gain for performing a k-th Kalman filter on the target physiological parameter detection value is determined.

[0072] In the disclosed embodiment, the Kalman filter gain of the kth Kalman filter on the target physiological parameter detection value is determined based on the estimation error and prediction error of the k-1th Kalman filter on the target physiological parameter detection value.

[0073] In an example of the present disclosure, initial values ​​are set for obtaining target physiological parameter values ​​using the Kalman filter algorithm. The initial value of the parameter k of the algorithm is pre-set to 0, the initial value of the target physiological parameter estimation error and the initial value of the target physiological parameter prediction error are the cross-validation root mean square error in the process of building the target physiological parameter detection model, and the initial value of the target physiological parameter estimation value output by the model the most recently is a preset value. All of the above values ​​change with the iteration of the Kalman filter algorithm.

[0074] In one implementation, the following calculation formula is used to calculate the Kalman filter gain K corresponding to the target physiological parameter detection value output by the model for the kth time.

[0075] K=e k-1 / (e k-1 +e j )

[0076] In the above formula, e k-1 is the k-1th estimation error, e j is the prediction error, k is a positive integer. Among them, the k-1th estimation error is the cross-validation root mean square error in the process of the k-1th target physiological parameter detection model outputting the target physiological parameter detection value. The prediction error is the cross-validation root mean square error in the process of the current target physiological parameter detection model outputting the target physiological parameter detection value.

[0077] In step S22, the k-th target physiological parameter detection estimation value is determined.

[0078] In the disclosed embodiment, the prediction error is updated based on the Kalman filter gain and the estimated error of the k-1th Kalman filter, and the kth target physiological parameter estimation value is determined based on the Kalman filter gain, the target physiological parameter detection value and the target physiological parameter estimation value of the k-1th Kalman filter.

[0079] In the embodiment of the present disclosure, the following formula is used to calculate the k-th target physiological parameter estimation value:

[0080] X k =X k-1 +K k ×(Z k -X k-1 )

[0081] In the above formula, X k-1 is the k-1th target physiological parameter estimation value, Kk is the kth Kalman filter gain, Z k It is the detection value of the target physiological parameter.

[0082] In one implementation, the following calculation formula is used to calculate the estimated error:

[0083] e k =(1-K)×e k-1

[0084] In the above formula, K is the Kalman gain, e k-1 is the estimated error of the k-1th Kalman filter on the target physiological parameter detection value.

[0085] In step S23, the k-th estimated value of the target physiological parameter is used as the target physiological parameter value.

[0086] In the disclosed embodiment, based on the k-1th target physiological parameter estimation error, the kth target physiological parameter estimation value, and the target physiological parameter detection value in the process of obtaining the target physiological parameter detection value, the target physiological parameter estimation value is assigned as the target physiological parameter value.

[0087] The physiological parameter detection method provided in the embodiment of the present disclosure collects the user's physiological state parameter set through a portable device and inputs it into a pre-trained physiological parameter detection model to obtain the target physiological parameter detection value. The result obtained by the model is subjected to Kalman filtering to obtain the final human body's blood oxygen saturation value. In this way, the problem that the blood oxygen saturation value cannot be tested in a state of motion and it is inconvenient to carry can be solved.

[0088] Figure 4 FIG. 1 is a flow chart showing a method of obtaining a target physiological parameter value by using a Kalman filter algorithm to obtain a target physiological parameter value according to an exemplary embodiment. Figure 4 As shown, the following steps are included.

[0089] Among them, in the embodiment of the present disclosure, the blood oxygen saturation value can be used as the target physiological parameter detection value.

[0090] In step S31, start.

[0091] In step S32, the parameters in the preset algorithm are initialized.

[0092] In step S33, the kth blood oxygen saturation value is obtained.

[0093] In the embodiment of the present disclosure, obtaining the kth blood oxygen saturation value can be understood as obtaining the kth blood oxygen saturation detection value output by the target physiological parameter detection model for the kth time.

[0094] In step S34, the Kalman gain corresponding to the k-th blood oxygen saturation value is calculated.

[0095] In the embodiment of the present disclosure, K=e k-1 / (e k-1 +e j ) calculation formula to obtain the Kalman gain corresponding to the k-th second blood oxygen saturation value.

[0096] In step S35, the k-th estimated blood oxygen saturation value is calculated.

[0097] In the embodiment of the present disclosure, by X k =X k-1 +K×(Z k -X k-1 ) calculation formula to obtain the k-th estimate.

[0098] In step S36, the estimated error is updated.

[0099] In the embodiment of the present disclosure, by e k =(1-K)×e k-1 Calculate the formula and update the k-th estimation error.

[0100] In step S37, a value is assigned.

[0101] In the disclosed embodiment, the k-th estimation error is assigned to the latest target physiological parameter value.

[0102] In step S38, the device is reset.

[0103] In step S39, end.

[0104] In the disclosed embodiment, the algorithm is terminated after the reset, and the latest target physiological parameter value is the target physiological parameter value. If the reset is not performed, the process returns to step S33.

[0105] Figure 5 FIG. 1 is a flow chart of a training target physiological parameter detection model according to an exemplary embodiment. Figure 5 As shown, the following steps are included.

[0106] In step S41, a first training sample is obtained.

[0107] In the embodiment of the present disclosure, the first training sample includes physiological state sample parameters of multiple different dimensions included in the physiological state parameter set.

[0108] In one implementation, the first training sample is a pulse perfusion index, a human heart rate value, a human body temperature value, a movement speed, a blood oxygen saturation value in a calm state, and a measured blood oxygen saturation value.

[0109] In one embodiment, the first training sample is obtained from a sensor in the wearable device, for example, the user wears a Bluetooth headset, and the training sample is obtained in response to the user being in motion or in a non-motion state.

[0110] In step S42, normalization is performed on physiological state sample parameters of multiple different dimensions to obtain second training samples.

[0111] In the embodiment of the present disclosure, the physiological state samples with different dimensions are subjected to maximum and minimum normalization processing, which can be understood as pre-processing the samples to eliminate the correlation between different features by converting the features of each dimension to the same value range.

[0112] In step S43, based on the second training sample, a target physiological parameter detection model is constructed using the radial basis function as the kernel function.

[0113] In the disclosed embodiment, a target physiological parameter detection model is constructed. Based on the normalized second training sample, the target physiological parameter detection model is constructed by using a support vector machine regression algorithm and a non-destructive testing model parameter optimization method with cross-validation root mean square as the target.

[0114] In one embodiment, the support vector machine regression algorithm contains a kernel function for mapping data into a high-dimensional space, and a polynomial is implicitly applied. The support vector regression can fit nonlinear trends well.

[0115] In one example, the kernel function contained in the support vector machine regression algorithm is a radial kernel function. It can be understood that no matter whether the sample data is high-dimensional or low-dimensional, and whether the data volume is large or small, the radial kernel function can show good classification performance.

[0116] In one embodiment, the expression of the radial basis kernel function is: K(X,X P )=exp[-(||XX P || 2 ) / 2g 2 ], where X P is the center of the kernel function, X is the sample data of the input model, ||XX P || 2 is the squared Euclidean distance between two eigenvectors, and g is the parameter of the kernel function.

[0117] In the disclosed embodiment, a first training sample is obtained based on a portable device to facilitate real-time detection of the physiological state parameter value of the user. The first training sample obtained is normalized to facilitate conversion of variables of different dimensions to the same value interval and eliminate the correlation between different features. Based on the normalized training sample, a target physiological parameter detection model is constructed using a radial basis kernel function, which can realize real-time monitoring of the physiological parameter values ​​of the human body in motion or non-motion state.

[0118] In the disclosed embodiment, the target physiological parameter detection model is constructed in two parts, using a support vector machine regression algorithm and a cross-validation root mean square error as the goal to find the parameters required for constructing a physiological parameter detection model.

[0119] Figure 6 is a flowchart showing a method of constructing a target physiological parameter detection model based on a second training sample and taking a radial basis function as a kernel function according to an exemplary embodiment. Figure 6 As shown, the following steps are included.

[0120] In step S51, a target physiological parameter detection prediction model is trained based on N groups of second training samples to obtain N cross-validation root mean square errors.

[0121] In the disclosed embodiment, a target physiological parameter detection prediction model is trained based on N groups of second training samples to obtain N corresponding cross-validation root mean square errors.

[0122] Among them, in the embodiment of the present disclosure, the target physiological parameter detection model is constructed with the radial basis function as the kernel function, and the target physiological parameter detection model includes a penalty factor and a kernel function parameter. Therefore, N cross-validation root mean square errors correspond to N groups of penalty factors and kernel function parameters, and N groups of penalty factors and kernel function parameters correspond to N target physiological parameter detection prediction models.

[0123] In step S52, based on grid optimization, the minimum cross-validation root mean square error is determined among N cross-validation root mean square errors, and a set of penalty factors and kernel function parameters corresponding to the minimum cross-validation root mean square error are determined as target penalty factors and target kernel function parameters.

[0124] In the embodiment of the present disclosure, a grid optimization method is adopted to determine the minimum cross-validation root mean square error from N cross-validation root mean square errors. When the cross-validation root mean square error is the minimum cross-validation root mean square error, the corresponding penalty factor and kernel function parameters are the optimal set, namely the target penalty factor and target kernel function parameters.

[0125] In one embodiment, the minimum cross-validation root mean square error is determined. It can be understood that: in the case of initial parameters, each set of penalty factors and kernel function parameters corresponds to a target physiological detection model and a cross-validation root mean square error. Therefore, the target penalty factor and target kernel function parameters are determined in a grid optimization manner. That is, the minimum cross-validation root mean square error is determined in a grid optimization manner.

[0126] Among them, in the embodiment of the present disclosure, a set of penalty factors and kernel function parameters corresponding to the minimum cross-validation root mean square error are the target penalty factors and target kernel function parameters. In other words, this set of penalty factors and kernel function parameters are the optimal penalty factors and optimal kernel function parameters for constructing the target physiological parameter detection model.

[0127] In step S53, a target physiological parameter detection prediction model corresponding to the target penalty factor and the target kernel function parameter is selected as the target physiological parameter detection model.

[0128] In the disclosed embodiment, the minimum cross-validation root mean square error is determined by grid optimization, and the target physiological parameter detection model is constructed using the kernel function parameters and penalty factors corresponding to the minimum cross-validation root mean square error.

[0129] A target physiological parameter detection method provided in an embodiment of the present disclosure, by pre-building a target physiological parameter detection model and embedding the model into a wearable device, can enable a user to monitor the condition of his or her body in real time in motion or non-motion state, can accurately know the target physiological parameter value in motion or non-motion state, and further, realize real-time monitoring of whether his or her own physiological function is healthy.

[0130] The following describes the method of determining the target penalty factor and the target kernel function parameters in the grid optimization method involved in the embodiment of the present disclosure. Figure 7 The present invention is a flowchart showing a method of determining a target penalty factor and target kernel function parameters in a grid optimization manner according to an exemplary embodiment.

[0131] like Figure 7 As shown, the parameters required to construct the target physiological parameter detection model are initialized. Among them, the maximum penalty factor Cmax, the maximum kernel function parameter gmax, the minimum penalty factor Cmin, the minimum kernel function parameter gmin, the penalty factor change step size Cstep, and the kernel function parameter change step size gstep are initialized. Initialize the optimal penalty factor bestC=0, the optimal kernel function parameter bestg=0, the minimum cross-validation root mean square error minRMSE=10, and the cross-validation fold fold=10. Given the penalty factor C=Cmin, the kernel function parameter g=gmin.

[0132] Among them, the parameters required for initializing and constructing the target physiological parameter detection model include a preset minimum cross-validation root mean square error, a preset step range of kernel function parameters, and a preset step range of penalty factors;

[0133] Based on the support vector machine regression algorithm, N cross-validation root mean square errors corresponding to N groups of training samples are determined. At the same time, the N cross-validation root mean square errors correspond to N groups of penalty factors and kernel function parameters, corresponding to N target physiological detection prediction models.

[0134] Determine whether the cross-validation root mean square error is less than the preset minimum cross-validation root mean square error. If the Nth cross-validation root mean square error is less than the preset minimum cross-validation root mean square error, assign the Nth cross-validation root mean square error to the preset minimum cross-validation root mean square error. At this time, the penalty factor and kernel function parameters corresponding to the Nth cross-validation root mean square error are the optimal penalty factor and kernel function parameters for constructing the target physiological parameter detection model.

[0135] Move the first step length of the kernel function parameter corresponding to the Nth cross-validation root mean square to the position of the second step length of the kernel function parameter, and determine whether the second step length of the kernel function parameter is within the range of the preset step length of the kernel function parameter.

[0136] If it is within the range of the preset kernel function parameter step size, then the N+1th cross-validation root mean square error is obtained again, and it is determined whether the N+1th cross-validation root mean square error is less than the new minimum cross-validation root mean square error.

[0137] If it is not within the range of the preset step size of the kernel function parameter, then move the first step size of the penalty factor corresponding to the N+1th cross-validation root mean square error to the second step size of the penalty factor, and the step size of the kernel function parameter is the minimum step size, and determine whether the second step size of the penalty factor is within the range of the preset step size of the penalty factor. The second step size of the penalty factor corresponds to the N+2th cross-validation root mean square error.

[0138] If the second step length of the penalty factor is within the range of the preset step length of the penalty factor, the cross-validation root mean square error corresponding to the second step length of the penalty factor is re-obtained, and it is determined whether the cross-validation root mean square error corresponding to the second step length of the penalty factor is less than the new minimum cross-validation root mean square error.

[0139] If the second step length of the penalty factor is not within the range of the preset step length of the penalty factor, the optimization process is terminated, and the penalty factor and kernel function parameters corresponding to the current minimum cross-validation root mean square error are selected.

[0140] If the Nth cross-validation root mean square error is greater than the preset minimum cross-validation root mean square error, move the first step length of the kernel function parameter corresponding to the Nth cross-validation root mean square error to the second step length of the kernel function parameter, and determine whether the second step length of the kernel function parameter is within the range of the preset kernel function parameter step length.

[0141] If it is within the range of the preset kernel function parameter step length, the corresponding N+1th cross-validation root mean square error is re-obtained, and it is determined whether the N+1th cross-validation root mean square error is less than the preset minimum cross-validation root mean square error.

[0142] If it is not within the range of the preset step size of the kernel function parameter, then move the first step size of the penalty factor corresponding to the N+1th cross-validation root mean square error to the second step size of the penalty factor, and the step size of the kernel function parameter is the minimum step size, and determine whether the second step size of the penalty factor is within the range of the preset step size of the penalty factor. The second step size of the penalty factor corresponds to the N+2th cross-validation root mean square error.

[0143] If the second step length of the penalty factor is within the range of the preset step length of the penalty factor, the cross-validation root mean square error corresponding to the second step length of the penalty factor is re-obtained, and it is determined whether the cross-validation root mean square error corresponding to the second step length of the penalty factor is less than the preset minimum cross-validation root mean square error.

[0144] If the second step length of the penalty factor is not within the range of the preset penalty factor step length, the optimization process is terminated and the current minimum cross-validation root mean square error is selected.

[0145] The grid-based optimization method provided in the embodiment of the present disclosure determines the minimum cross-validation root mean square among N cross-validation root mean square errors, and its corresponding penalty factor and kernel function parameters are the optimal penalty factor and kernel function parameters. By using the grid-based optimization method, the optimal penalty factor and kernel function parameters can be found globally, and the target physiological parameter detection model constructed based on the penalty factor and kernel function parameters can be the optimal model, thereby ensuring the accuracy of the target physiological parameter value.

[0146] In the disclosed embodiment, the wearable device may include a sensor group, a processor, and a memory, wherein the sensor group may be used to obtain physiological state parameters in multiple dimensions.

[0147] Among them, the physiological state parameters of multiple dimensions are used to determine the target physiological parameter values ​​to be detected, and the target physiological parameter values ​​include the physiological parameter values ​​of the user in a motion state and / or a non-motion state.

[0148] The processor may be used to determine a target physiological parameter value that matches the physiological state parameter.

[0149] In the disclosed embodiment, the target physiological parameter to be detected can be determined by a sensor group of a wearable device, for example, a sensor group installed in a Bluetooth headset. The target physiological parameter can be a physiological parameter of the user in motion, or the target physiological parameter can be a physiological parameter of the user in a non-motion state.

[0150] In the disclosed embodiment, the sensor group and the processor are connected in communication, so that the sensor group can send the physiological state parameters of multiple dimensions acquired by the sensor group to the processor, and the processor can determine the target physiological parameter value based on the physiological state parameters. The processor can be a processor located in a wearable device, or can be a processor located in other electronic devices.

[0151] In the embodiment of the present disclosure, the wearable device may further include a communication component. The multiple physiological state parameters acquired by the sensor group are sent to other electronic devices through the communication component in the wearable device. For example, the physiological state parameters acquired by the wearable device may be sent to an electronic device connected to the wearable device via Bluetooth via Bluetooth.

[0152] In the disclosed embodiment, each sensor in the wearable device sensor group can be connected to a processor, and the sensor can transmit corresponding physiological state parameter information to the processor, and the processor of the wearable device processes the physiological state parameters to obtain corresponding target physiological state parameters.

[0153] In the disclosed embodiment, the wearable device can be connected to an electronic device. After the wearable device obtains physiological state parameters of multiple dimensions through a sensor group, the physiological state parameters of multiple dimensions can be sent to the electronic device connected to the wearable device. For example, the obtained physiological state parameters can be sent to the electronic device via Bluetooth, and the electronic device receives the multi-dimensional physiological state parameters and determines the target physiological state parameters.

[0154] In the embodiment of the present disclosure, the heart rate parameter value can be obtained through the heart rate sensor in the sensor group. When the wearable device detects that the user's heart rate needs to be measured, the processor calls the heart rate sensor to test the user's current heart rate, so that the user's heart rate can be obtained as a dimension in the physiological state parameter.

[0155] In the disclosed embodiment, the temperature parameter value can be obtained through the temperature sensor in the sensor group. When the wearable device determines that the user's body temperature needs to be measured, the temperature of the user is measured based on the temperature sensor through the processor temperature sensor, and the measured temperature value is used as a dimension in the physiological state parameter. In the disclosed embodiment, the speed parameter value can be obtained through the acceleration sensor in the sensor group. When the wearable device determines that the user's current motion state needs to be measured, the acceleration sensor is called by the processor, and the user's current motion state is measured based on the acceleration sensor, and the measured speed value is used as a dimension in the physiological state parameter, so that the wearable device can accurately detect the user's current speed parameter during the user's motion.

[0156] In the disclosed embodiment, the pulse perfusion index parameter can be obtained through the short-wave infrared sensor in the sensor group. When the wearable device determines that the user's pulse perfusion index needs to be obtained, the short-wave infrared sensor is called, wherein the infrared short-wave sensor can emit 660 nanometers and 940 nanometers near-infrared light, based on the diffuse reflection of infrared light on the user's skin, receive diffuse reflection spectrum signals, and determine the corresponding pulse perfusion index physiological state parameters based on the diffuse reflection spectrum signals.

[0157] In one example, near infrared light emitted by a short-wave infrared sensor can be used as an incident light source to determine the blood oxygen saturation value, wherein the blood oxygen saturation value can be the blood oxygen saturation value of the user in a calm state, or can be the blood oxygen saturation value of the user in the current state.

[0158] In the disclosed embodiment, the pulse perfusion index parameter and the blood oxygen saturation parameter can be simultaneously determined by the spectral signal acquired by the short-wave infrared sensor.

[0159] In the embodiments of the present disclosure, one or more sensors in the sensor group may respond to instructions issued by the wearable device at the same time, or one or more sensors in the sensor group may respond to instructions issued by the wearable device in sequence, without limitation herein.

[0160] In the embodiment of the present disclosure, the wearable device may be a headset, for example, a Bluetooth headset.

[0161] In the disclosed embodiment, a target physiological parameter detection model is embedded into a wearable device to achieve detection of physiological state parameters based on the wearable device worn by the user in a sports scenario.

[0162] Figure 8 A scene diagram of detecting a user's physiological parameters based on a wearable device worn by the user is shown. Figure 8As shown in FIG, the user wears a Bluetooth headset while exercising. The Bluetooth headset contains a trained target physiological parameter detection model, such as Figure 8 As shown. Based on the physiological state parameter set obtained by the Bluetooth headset, the processed physiological state parameter set is input into the target physiological parameter detection model, and the Kalman filter is performed based on the output result to obtain the user's physiological parameter value. The disclosed embodiment provides a physiological parameter detection method, which obtains the user's physiological state parameter set through a wearable device, and inputs the acquired physiological state parameter set into a predetermined target physiological parameter detection model to obtain the target physiological parameter value. Therefore, based on the wearable device worn by the user, in which the target physiological parameter detection model is embedded in the device, there is no need to remain calm during the detection process, and it is easy to carry, and can monitor the physiological parameter values ​​of the human body in real time when the user is in motion. Furthermore, in this way, the process of applying physiological parameter detection in portable devices can be accelerated, providing a theoretical basis for the application of physiological parameter detection technology in wireless Bluetooth headsets.

[0163] Based on the same concept, the embodiment of the present disclosure also provides a device for detecting physiological parameters.

[0164] It is understandable that the device for detecting physiological parameters provided by the embodiments of the present disclosure includes hardware structures and / or software modules corresponding to the execution of each function in order to realize the above functions. In combination with the units and algorithm steps of each example disclosed in the embodiments of the present disclosure, the embodiments of the present disclosure can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art may use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the technical solution of the embodiments of the present disclosure.

[0165] Fig. 9 is a block diagram of a device for detecting physiological parameters according to an exemplary embodiment. Fig. 9 The physiological parameter detection device 100 includes an acquisition unit 101 and a processing unit 102 .

[0166] The acquisition unit 101 is used to obtain a set of physiological state parameters of the user, wherein the set of physiological state parameters includes physiological state parameters of multiple dimensions, and the physiological state parameters of multiple dimensions are used to determine target physiological parameter values ​​to be detected, and the target physiological parameter values ​​include physiological parameter values ​​of the user in a motion state and / or a non-motion state.

[0167] The processing unit 102 is used to determine a target physiological parameter value, wherein the target physiological parameter value is determined according to a normalized physiological state parameter and a target physiological parameter detection model, wherein the normalized physiological state parameter is obtained by normalizing multiple physiological state parameters in a physiological state parameter set.

[0168] In one embodiment, the processing unit 102 determines the target physiological parameter value according to the normalized physiological state parameter and the target physiological parameter detection model in the following manner: Kalman filtering is performed on the target physiological parameter detection value corresponding to the output result to obtain the target physiological parameter value. The output result is the output result corresponding to the normalized physiological state parameter of the target physiological parameter detection model.

[0169] In one embodiment, the processing unit 102 performs Kalman filtering on the target physiological parameter detection value corresponding to the output result to obtain the target physiological parameter value in the following manner: based on the estimation error and prediction error of the k-1th Kalman filtering of the target physiological parameter detection value, the Kalman filter gain of the k-1th Kalman filtering of the target physiological parameter detection value is determined. Based on the Kalman filter gain and the estimation error of the k-1th Kalman filtering, the prediction error is updated, and based on the Kalman filter gain, the target physiological parameter detection value, and the target physiological parameter estimation value of the k-1th Kalman filtering, the k-th target physiological parameter estimation value is determined. The k-th target physiological parameter estimation value is used as the target physiological parameter value.

[0170] In one embodiment, the processing unit 102 is also used to determine a target physiological parameter detection model. The processing unit determines the target physiological parameter detection model in the following manner: obtain a first training sample, the first training sample includes physiological state sample parameters of multiple dimensions included in the physiological state parameter set. Normalize the physiological state sample parameters of multiple dimensions to obtain a second training sample. Based on the second training sample, a target physiological parameter detection model is constructed using a radial basis function as a kernel function.

[0171] In one embodiment, the processing unit 102 is further configured to construct a target physiological parameter detection model based on the second training sample and using the radial basis function as the kernel function: the target physiological parameter detection prediction model is trained based on N groups of second training samples to obtain N cross-validation root mean square errors, the N cross-validation root mean square errors correspond to N groups of penalty factors and kernel function parameters, and the N groups of penalty factors and kernel function parameters correspond to N target physiological parameter detection prediction models. Based on grid optimization, the minimum cross-validation root mean square error is determined among the N cross-validation root mean square errors, and a set of penalty factors and kernel function parameters corresponding to the minimum cross-validation root mean square error are determined as the target penalty factor and target kernel function parameters. The target physiological parameter detection prediction model corresponding to the target penalty factor and target kernel function parameters is selected as the target physiological parameter detection model.

[0172] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0173] Fig.10 2 is a block diagram of an apparatus 200 for physiological parameter detection according to an exemplary embodiment. For example, the apparatus 200 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

[0174] Reference Fig.10 , the device 200 may include one or more of the following components: a processing component 202 , a memory 204 , a power component 206 , a multimedia component 208 , an audio component 210 , an input / output (I / O) interface 212 , a sensor component 214 , and a communication component 216 .

[0175] The processing component 202 generally controls the overall operation of the device 200, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 202 may include one or more processors 220 to execute instructions to perform all or part of the steps of the above-described method. In addition, the processing component 202 may include one or more modules to facilitate interaction between the processing component 202 and other components. For example, the processing component 202 may include a multimedia module to facilitate interaction between the multimedia component 208 and the processing component 202.

[0176] The memory 204 is configured to store various types of data to support operations on the device 200. Examples of such data include instructions for any application or method operating on the device 200, contact data, phone book data, messages, pictures, videos, etc. The memory 204 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0177] The power component 206 provides power to the various components of the device 200. The power component 206 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device 200.

[0178] The multimedia component 208 includes a screen that provides an output interface between the device 200 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 208 includes a front camera and / or a rear camera. When the device 200 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

[0179] The audio component 210 is configured to output and / or input audio signals. For example, the audio component 210 includes a microphone (MIC), and when the device 200 is in an operation mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 204 or sent via the communication component 216. In some embodiments, the audio component 210 also includes a speaker for outputting audio signals.

[0180] I / O interface 212 provides an interface between processing component 202 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: a home button, a volume button, a start button, and a lock button.

[0181] The sensor assembly 214 includes one or more sensors for providing various aspects of the status assessment of the device 200. For example, the sensor assembly 214 can detect the open / closed state of the device 200, the relative positioning of components, such as the display and keypad of the device 200, the sensor assembly 214 can also detect the position change of the device 200 or a component of the device 200, the presence or absence of user contact with the device 200, the orientation or acceleration / deceleration of the device 200 and the temperature change of the device 200. The sensor assembly 214 can include a proximity sensor configured to detect the presence of a nearby object without any physical contact. The sensor assembly 214 can also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 214 can also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor or a temperature sensor.

[0182] The communication component 216 is configured to facilitate wired or wireless communication between the device 200 and other devices. The device 200 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 216 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 216 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0183] In an exemplary embodiment, the apparatus 200 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0184] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 204 including instructions, and the instructions can be executed by the processor 220 of the device 200 to perform the above method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0185] It is to be understood that in the present disclosure, "plurality" refers to two or more than two, and other quantifiers are similar. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. The singular forms "a", "the" and "the" are also intended to include plural forms, unless the context clearly indicates other meanings.

[0186] It is further understood that the terms "first", "second", etc. are used to describe various information, but such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other, and do not indicate a specific order or degree of importance. In fact, the expressions "first", "second", etc. can be used interchangeably. For example, without departing from the scope of the present disclosure, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as the first information.

[0187] It can be further understood that, unless otherwise specified, “connection” includes a direct connection without other components between the two, and also includes an indirect connection with other components between the two.

[0188] It is further understood that, although the operations are described in a specific order in the drawings in the embodiments of the present disclosure, it should not be understood as requiring the operations to be performed in the specific order shown or in a serial order, or requiring the execution of all the operations shown to obtain the desired results. In certain environments, multitasking and parallel processing may be advantageous.

[0189] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any modifications, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary technical means in the art that are not disclosed in the present disclosure.

[0190] It should be understood that the present disclosure is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the scope of the appended claims.

Claims

1. A physiological parameter detection method, characterized in that: The method comprises: Acquire a physiological state parameter set of the user, wherein the physiological state parameter set includes physiological state parameters of multiple dimensions, and the physiological state parameters of the multiple dimensions are used to determine target physiological parameter values ​​to be detected, wherein the target physiological parameter values ​​include physiological parameter values ​​of the user in a motion state and / or a non-motion state; Determine a target physiological parameter value, wherein the target physiological parameter value is determined according to a normalized physiological state parameter and a target physiological parameter detection model, wherein the normalized physiological state parameter is obtained by normalizing the multiple physiological state parameters in the physiological state parameter set.

2. The method according to claim 1, characterized in that: The target physiological parameter value is determined according to the normalized physiological state parameter and the target physiological parameter detection model in the following manner: Performing Kalman filtering on the target physiological parameter detection value corresponding to the output result to obtain the target physiological parameter value; The output result is the output result corresponding to the normalized physiological state parameter whose input is the target physiological parameter detection model.

3. The method according to claim 2, characterized in that The step of performing Kalman filtering on the target physiological parameter detection value corresponding to the output result to obtain the target physiological parameter value includes: Determining a Kalman filter gain of a k-th Kalman filter performed on the target physiological parameter detection value based on an estimation error and a prediction error of a k-1-th Kalman filter performed on the target physiological parameter detection value; Based on the Kalman filter gain and the estimated error of the k-1th Kalman filter, the prediction error is updated, and based on the Kalman filter gain, the target physiological parameter detection value and the target physiological parameter estimated value of the k-1th Kalman filter, the kth target physiological parameter estimated value is determined; The k-th target physiological parameter estimation value is used as the target physiological parameter value.

4. The method according to any one of claims 1 to 3, characterized in that The target physiological parameter detection model is determined in the following manner: Acquire a first training sample, where the first training sample includes physiological state sample parameters of multiple dimensions included in the physiological state parameter set; Normalizing the physiological state sample parameters of the multiple dimensions to obtain a second training sample; Based on the second training sample, the target physiological parameter detection model is constructed using the radial basis function as the kernel function.

5. The method according to claim 4, characterized in that The step of constructing the target physiological parameter detection model based on the second training sample and taking the radial basis function as the kernel function includes: Training the target physiological parameter detection prediction model based on N groups of the second training samples to obtain N cross-validation root mean square errors, wherein the N cross-validation root mean square errors correspond to N groups of penalty factors and kernel function parameters, and the N groups of penalty factors and kernel function parameters correspond to N target physiological parameter detection prediction models; Based on grid optimization, a minimum cross-validation root mean square error is determined among the N cross-validation root mean square errors, and a set of penalty factors and kernel function parameters corresponding to the minimum cross-validation root mean square error are determined as target penalty factors and target kernel function parameters; A target physiological parameter detection prediction model corresponding to the target penalty factor and the target kernel function parameter is selected as the target physiological parameter detection model.

6. A device for detecting physiological parameters, characterized in that: The device comprises: an acquisition unit, configured to acquire a physiological state parameter set of a user, wherein the physiological state parameter set includes physiological state parameters of multiple dimensions, and the physiological state parameters of multiple dimensions are used to determine target physiological parameter values ​​to be detected, wherein the target physiological parameter values ​​include physiological parameter values ​​of the user in a motion state and / or a non-motion state; A processing unit is used to determine a target physiological parameter value, wherein the target physiological parameter value is determined based on a normalized physiological state parameter and a target physiological parameter detection model, wherein the normalized physiological state parameter is obtained by normalizing the multiple physiological state parameters in the physiological state parameter set.

7. The device according to claim 6, characterized in that: The processing unit determines the target physiological parameter value according to the normalized physiological state parameter and the target physiological parameter detection model in the following manner: Performing Kalman filtering on the target physiological parameter detection value corresponding to the output result to obtain the target physiological parameter value; The output result is the output result corresponding to the normalized physiological state parameter whose input is the target physiological parameter detection model.

8. The device according to claim 7, characterized in that: The processing unit performs Kalman filtering on the target physiological parameter detection value corresponding to the output result to obtain the target physiological parameter value in the following manner: Determining a Kalman filter gain of a k-th Kalman filter performed on the target physiological parameter detection value based on an estimation error and a prediction error of a k-1-th Kalman filter performed on the target physiological parameter detection value; Based on the Kalman filter gain and the estimated error of the k-1th Kalman filter, the prediction error is updated, and based on the Kalman filter gain, the target physiological parameter detection value and the target physiological parameter estimated value of the k-1th Kalman filter, the kth target physiological parameter estimated value is determined; The k-th target physiological parameter estimation value is used as the target physiological parameter value.

9. The device according to any one of claims 6 to 8, characterized in that The processing unit is also used to determine the target physiological parameter detection model; The processing unit determines the target physiological parameter detection model in the following manner: Acquire a first training sample, where the first training sample includes physiological state sample parameters of multiple dimensions included in the physiological state parameter set; Normalizing the physiological state sample parameters of the multiple dimensions to obtain a second training sample; Based on the second training sample, the target physiological parameter detection model is constructed using the radial basis function as the kernel function.

10. The device according to claim 9, characterized in that: The processing unit is further configured to construct the target physiological parameter detection model based on the second training sample and using the radial basis function as a kernel function: Training the target physiological parameter detection prediction model based on N groups of the second training samples to obtain N cross-validation root mean square errors, wherein the N cross-validation root mean square errors correspond to N groups of penalty factors and kernel function parameters, and the N groups of penalty factors and kernel function parameters correspond to N target physiological parameter detection prediction models; Based on grid optimization, a minimum cross-validation root mean square error is determined among the N cross-validation root mean square errors, and a set of penalty factors and kernel function parameters corresponding to the minimum cross-validation root mean square error are determined as target penalty factors and target kernel function parameters; A target physiological parameter detection prediction model corresponding to the target penalty factor and the target kernel function parameter is selected as the target physiological parameter detection model.

11. A wearable device, characterized in that: include: A sensor group, used for acquiring physiological state parameters of multiple dimensions, wherein the physiological state parameters of multiple dimensions are used for determining target physiological parameter values ​​to be detected, wherein the target physiological parameter values ​​include physiological parameter values ​​of the user in a motion state and / or a non-motion state; processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to: determine a target physiological parameter value matching the physiological state parameter, and the target physiological parameter value is obtained by executing the physiological detection method described in any one of claims 1-5.

12. The wearable device according to claim 11, characterized in that: The sensor group is communicatively connected to the processor to send the physiological state parameters of multiple dimensions acquired by the sensor group to the processor; The processor determines a target physiological parameter value based on the physiological state parameter.

13. The wearable device according to claim 11, characterized in that: The wearable device further comprises: a communication component; The communication component is used to send the physiological state parameter to an electronic device different from the wearable device, receive a target physiological parameter value determined by the electronic device based on the physiological state parameter, and send the target physiological parameter value to the processor.

14. The wearable device according to claim 11, characterized in that: The sensor group includes one or more of the following sensors: Short-wave infrared sensor; Heart rate sensor; temperature sensor; and Accelerometer.

15. The wearable device according to claim 11, characterized in that: The wearable device is: headphones.

16. A storage medium, characterized in that: The storage medium stores instructions, and when the instructions in the storage medium are executed by a processor of the terminal, the terminal is enabled to execute the physiological parameter detection method described in any one of claims 1-5.

Images