A method and device for denoising exercise pulse waves

Abstract

The present invention designs a method and device for denoising exercise pulse waves, including the following steps: acquiring the currently collected sensor signal and the N-axis accelerometer signal, and respectively calculating the correlation coefficient between the sensor signal and the accelerometer signals of each axis as the first correlation coefficient; where N is a positive integer; obtaining the optimal step factor as the first optimal step factor according to the relational expression of the curve fitting of each correlation coefficient and each step factor; filtering the sensor signal according to the first optimal step factor and a preset adaptive filtering algorithm to obtain the denoised pulse wave signal. The present invention can accurately describe the relationship between the correlation coefficient and the step factor through the relational expression obtained by curve fitting, and can accurately find the optimal step factor through the relational expression in different exercise scenarios, thereby improving the denoising accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of digital signal processing, and particularly to a method and device for denoising motion pulse waves. Background Art

[0002] Adaptive filtering is a commonly used filtering method for obtaining heart rate by denoising pulse waves. It includes a parameter-adjustable digital filter (Adaptive filter) and an adaptive filtering algorithm (Adaptive Algorithm). The input signal generates an output signal through the parameter-adjustable digital filter. The desired signal is compared with the output signal to obtain an error signal. In order to make the error signal as small as possible, the filtering process needs to be adjusted through several cycles. Different adaptive filtering algorithms will result in different filtering results. Common adaptive filtering algorithms include: Least Mean Square (LMS) filtering algorithm, Recursive Least Square (RLS) filtering algorithm, and Normalized Least Mean Square (NLMS) adaptive filtering algorithm. The iterative process of the commonly used LMS algorithm includes:

[0003] Calculating the output signal generated by the input signal through the parameter-adjustable digital filter, y(n) = w T (n)x(n);

[0004] Calculating the error signal, e(n) = d(n) - y(n);

[0005] Updating the weight vector of the adaptive linear combiner at the next moment, w(n + 1) = w(n) + 2μe(n)x(n);

[0006] Wherein, x(n) is the input signal, d(n) and y(n) are the desired signal and the output signal respectively, e(n) is the error signal, w(n) and w(n + 1) are the weight vectors of the adaptive linear combiner at this moment and the next moment respectively, μ is the step factor, and the convergence condition of the LMS algorithm is to obtain the minimum mean square error (that is, the error between the desired signal and the output signal of the filter is infinitely reduced).

[0007] Initial convergence speed, time-varying system tracking ability, and steady-state offset are the three most important technical indicators for measuring the advantages and disadvantages of adaptive filtering algorithms. Due to the inevitable interference noise at the signal input end, the adaptive filtering algorithm will generate parameter offset noise. The greater the interference noise, the greater the resulting offset noise. Reducing the step size factor can reduce the parameter offset noise of the adaptive filtering algorithm and improve the convergence accuracy of the algorithm. However, reducing the step size factor will reduce the convergence speed and tracking speed of the algorithm. Therefore, the requirements for adjusting the step size factor of the adaptive filtering algorithm in terms of convergence speed, time-varying system tracking speed, and convergence accuracy are mutually contradictory.

[0008] Therefore, the stability of the filter can be improved by setting an appropriate step size factor. Existing technologies usually use a fixed value of the step size for adaptive filtering to remove motion artifacts. However, in the scenario of motion tests, the proportions of the pulse wave signal components and motion noise components in the sensor signals obtained under different motion states (walking, running, ball games, etc.) and different motion intensities are different. Therefore, when using adaptive filtering, the selected step size factor also needs to be adjusted according to the actual situation. The existing patent CN108652609A (A heart rate acquisition method, system, and wearable device) established the relationship between the magnitude of each motion signal, the motion state, and the magnitude of the step size factor, and gave the value ranges of the step size factor including the states of running, cycling, and walking.

[0009] However, the existing technologies consider limited motion states and cannot adapt to different motion scenarios; every time the motion scenario or motion intensity is changed, the value of the step size factor needs to be manually set; moreover, the range of the set step size factor is large and cannot give an accurate reference in practice; in addition, the signal intensities obtained by different types of accelerometer chips in the same environment may also bring numerical differences in the step size factor. Summary of the Invention

[0010] The embodiments of the present invention provide a method and device for removing motion pulse waves, which preprocess the step size factor by establishing a relationship formula for the optimal step size factor, can adapt to different motion scenarios, and adjust the step size factor in real time.

[0011] In a first aspect, the embodiments of the present invention provide a method for removing motion pulse waves, and the method includes:

[0012] Obtain the currently collected sensor signal and the N-axis accelerometer signal, and calculate the correlation coefficient between the sensor signal and each axis accelerometer signal as the first correlation coefficient; where N is a positive integer;

[0013] Obtain the optimal step size factor as the first optimal step size factor according to the relationship formula of curve fitting of each correlation coefficient and each step size factor;

[0014] Filter the sensor signal according to the first optimal step size factor and the preset adaptive filtering algorithm to obtain a denoised pulse wave signal.

[0015] In the present invention, by fitting the relational expression of the optimal step size factor, the step size factor is preprocessed, and the correlation coefficient can be calculated for the sensor signal and the N-axis accelerometer signal obtained in real time. According to the fitted relational expression, a more accurate optimal step size factor under the corresponding motion scenario can be obtained without manually modifying the step size factor, which can make the filtered spectrum data more accurate and reliable, thereby improving the accuracy and credibility of obtaining the exercise heart rate value.

[0016] Further, obtaining the optimal step size factor as the first optimal step size factor according to the fitting curve of each correlation coefficient and the step size factor includes:

[0017] Collect the sensor signals and N-axis accelerometer signals in M different motion scenarios; where M is a positive integer;

[0018] Calculate the correlation coefficient of each group of sensor signals and N-axis accelerometer signals as the second correlation coefficient;

[0019] According to the preset adaptive filtering algorithm, adjust the step size factors of M groups respectively, and select the step size factor with the highest waveform amplitude and the most obvious period after filtering as the second optimal step size factor under the current motion scenario to obtain the corresponding M groups of step size factors;

[0020] Perform curve fitting on the M groups of second correlation coefficients and the second optimal step size factor to obtain a relational expression;

[0021] Obtain the first optimal step size factor according to the first correlation coefficient and the relational expression.

[0022] In the present invention, by calculating the correlation coefficient of the sensor signal and the N-axis accelerometer signal in different scenarios, the proportion of the exercise intensity and the pulse wave / motion artifact in the sensor signal is described, and the information of the motion artifact is reflected by the signal of the accelerometer. The larger the correlation coefficient, the greater the exercise intensity and the greater the motion interference in the sensor signal, and a larger step size factor is required for denoising, which can more simply and conveniently reflect the value of the step size factor. By curve fitting, a relational expression is obtained to accurately describe the relationship between the correlation coefficient and the step size factor. Under different correlation coefficients, that is, in different motion scenarios, the optimal step size factor can be accurately found through the relational expression, thereby improving the denoising accuracy.

[0023] Further, calculating the correlation coefficient of each group of sensor signals and N-axis accelerometer signals as the second correlation coefficient includes:

[0024] Extract the data of the sensor signal and the N-axis accelerometer data with the same time period and the same length, and calculate the correlation coefficients between the data of the sensor signal and each axis data of the N-axis accelerometer respectively, and obtain N groups of correlation coefficients as the second correlation coefficient.

[0025] Preferably, the relational expression is specifically:

[0026] Or Wherein, is the correlation coefficient in the current motion scenario, μ * is the optimal step size factor in the current motion scenario, abs(·) and exp(·) are respectively the operations of taking the absolute value and the exponential function with the natural constant e as the base.

[0027] Preferably, the specific calculation formula of the correlation relationship is the Pearson correlation coefficient calculation formula:

[0028] Wherein, pw i and acc i (j) are respectively the i-th signal data of the sensor and the i-th signal data on the j-th axis of the N-axis accelerometer, 1 ≤ j ≤ N; and are respectively the mean value of the sensor signal data and the mean value of the signal data on the j-th axis of the N-axis accelerometer; L is the length of the extracted sensor signal data and N-axis accelerometer signal data; a(j) is the correlation coefficient between the sensor signal and the signal data on the j-th axis of the N-axis accelerometer; or,

[0029] Calculate using the Spearman correlation formula:

[0030] Wherein, is the rank difference between the i-th signal data of the sensor and the i-th signal data on the j-th (1 ≤ j ≤ N) axis of the N-axis sensor.

[0031] The present invention calculates by the Pearson correlation coefficient calculation method or the Spearman correlation coefficient calculation method, and uses the absolute value of the average value of the correlation coefficients to reflect the strength of the correlation between the sensor signal and the accelerometer signal. The larger the absolute value of the average value of the correlation coefficients, the stronger the correlation, which can more accurately describe the correlation relationship between the sensor signal and the accelerometer signal, obtain a credible correlation coefficient, and can make the curve fitting under different motion scenarios more scientific and reliable, thereby improving the denoising ability of adaptive filtering.

[0032] In a second aspect, an embodiment of the present invention provides a motion pulse wave denoising device, and the device includes:

[0033] The first correlation coefficient calculation module is used to obtain the currently collected sensor signal and the N-axis accelerometer signal, and calculate the correlation coefficient between the sensor signal and each axis accelerometer signal as the first correlation coefficient; where N is a positive integer;

[0034] The first optimal step size factor calculation module is used to obtain the optimal step size factor as the first optimal step size factor according to the curve fitting relationship between each correlation coefficient and each step size factor;

[0035] The pulse wave signal denoising module is used to filter the sensor signal according to the first optimal step size factor and a preset adaptive filtering algorithm to obtain the denoised pulse wave signal.

[0036] Furthermore, the first optimal step size factor calculation module is specifically:

[0037] Collect the sensor signal and the N-axis accelerometer signal under M groups of different motion scenarios; where M is a positive integer;

[0038] Calculate the correlation coefficient between each group of sensor signals and the N-axis accelerometer signal as the second correlation coefficient;

[0039] According to the preset adaptive filtering algorithm, adjust the step size factors of M groups respectively, and select the step size factor with the highest waveform amplitude and the most obvious period after filtering as the second optimal step size factor under the current motion scenario, and obtain the corresponding M groups of step size factors;

[0040] Perform curve fitting on the M groups of second correlation coefficients and the second optimal step size factor to obtain a relationship;

[0041] Obtain the first optimal step size factor according to the first correlation coefficient and the relationship.

[0042] Furthermore, the calculation of the correlation coefficient between each group of sensor signals and the N-axis accelerometer signal as the second correlation coefficient includes:

[0043] Extract the data of the sensor signal and the N-axis accelerometer data with the same time period and the same length, and calculate the correlation coefficient between the data of the sensor signal and each axis data of the N-axis accelerometer respectively, and obtain N groups of correlation coefficients as the second correlation coefficient.

[0044] Preferably, the relationship is specifically:

[0045] Or Where is the correlation coefficient under the current motion scenario, μ * is the optimal step size factor under the current motion scenario, abs(·) and exp(·) are respectively the absolute value taking and the exponential function operation with the natural constant e as the base.

[0046] Preferably, the specific calculation formula of the correlation coefficient is the Pearson correlation coefficient calculation formula:

[0047]

[0048] where pw i and acc i () are the i-th signal data of the sensor and the i-th signal data on the j-th axis of the N-axis accelerometer, respectively, where 1 ≤ j ≤ N; and are the mean values of the sensor's signal data and the signal data on the j-th axis of the N-axis accelerometer, respectively; L is the length of the extracted sensor signal data and N-axis accelerometer signal data; a(j) is the correlation coefficient between the sensor signal and the signal data on the j-th axis of the N-axis accelerometer; or,

[0049] Calculate using the Spearman correlation formula:

[0050]

[0051] where is the rank difference between the i-th signal data of the sensor and the i-th signal data on the j-th axis of the N-axis accelerometer.

[0052] Through the first correlation coefficient calculation module, the first optimal step size factor calculation module, and the pulse wave signal denoising module, the present invention calculates the correlation coefficient in the current motion scenario to obtain the optimal step size factor, and then denoises the sensor signal according to the preset adaptive filtering algorithm to obtain a pulse wave signal with less noise influence. There is no need to manually modify the step size factor according to different motion scenarios. By fitting the curve to obtain the relationship between the correlation coefficient and the step size factor, a more accurate step size factor can be obtained, which can adapt to different types and intensities of motion, thereby improving the denoising ability of the adaptive filtering algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is a schematic flowchart of a method for denoising motion pulse waves provided by an embodiment of the present invention;

[0054] Figure 2 is a schematic flowchart of curve fitting of a method for denoising motion pulse waves provided by an embodiment of the present invention;

[0055] Figure 3 is a schematic diagram of curve fitting of a method for denoising motion pulse waves provided by an embodiment of the present invention;

[0056] Figure 4 is a schematic flowchart of a method for denoising motion pulse waves provided by an embodiment of the present invention;

[0057] Figure 5 It is a schematic comparison diagram before and after denoising of a motion pulse wave provided by an embodiment of the present invention;

[0058] Figure 6 It is a schematic diagram of heart rate values under different motion scenarios after denoising of a motion pulse wave provided by an embodiment of the present invention;

[0059] Figure 7 It is a schematic structural diagram of a motion pulse wave denoising device provided by an embodiment of the present invention. Detailed implementation manners

[0060] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0061] Refer to Figure 1 , which is a motion pulse wave denoising method provided by an embodiment of the present invention, including steps S101 to S103, specifically as follows:

[0062] Step S101: Obtain the currently collected sensor signal and N-axis accelerometer signal, and calculate the correlation coefficient between the sensor signal and each axis accelerometer signal as the first correlation coefficient; where N is a positive integer.

[0063] In an optional implementation manner, after obtaining the currently collected sensor signal and N-axis accelerometer signal, the calculated first correlation coefficient may be the mean of the correlation coefficients between the sensor signal and each axis accelerometer signal, or the correlation coefficient between the mean of each axis accelerometer signal and the sensor signal, which is not limited herein; in addition, the N-axis accelerometer is not limited to a 3-axis accelerometer or a 6-axis accelerometer.

[0064] Step S102: Obtain the optimal step factor as the first optimal step factor according to the curve fitting relationship of each correlation coefficient and each step factor.

[0065] Specifically refer to Figure 2 , which is a flow schematic diagram of curve fitting of a motion pulse wave denoising method provided by an embodiment of the present invention, including steps S21 to S24, specifically as follows:

[0066] Step S21: Collect sensor signals and N-axis accelerometer signals under M different motion scenarios; where M is a positive integer.

[0067] Step S22: Calculate the correlation coefficient between each group of sensor signals and the N-axis accelerometer signals as the second correlation coefficient.

[0068] Specifically, extract the data of the sensor signals and the N-axis accelerometer data with the same time period and the same length, and calculate the correlation coefficient between the data of the sensor signals and each axis data of the N-axis accelerometer respectively, obtaining N groups of correlation coefficients as the second correlation coefficient.

[0069] It should be noted that each group of the calculated second correlation coefficients can be the mean value of the correlation coefficients between the sensor signals and the signals of each axis accelerometer, or the correlation coefficient between the mean value of the signals of each axis accelerometer and the sensor signals, which is not limited herein.

[0070] Step S23: According to the preset adaptive filtering algorithm, adjust the step size factors of M groups respectively, and select the step size factor with the highest waveform amplitude and the most obvious period after filtering as the second optimal step size factor in the current motion scenario, obtaining the corresponding M groups of step size factors.

[0071] Among them, the preset adaptive filtering algorithm can be the LMS algorithm, the RLS algorithm or the NLMS algorithm, which is not limited herein. As an example, the LMS algorithm is selected as the preset adaptive filtering algorithm. Adjust the step size factors of the LMS algorithm under M groups of correlation coefficients respectively, debug the step size factors for several times, and select the step size factor with the highest waveform amplitude and the most obvious period after filtering as the optimal step size factor in the current motion scenario.

[0072] It should be noted that since the correlation coefficient can reflect the exercise intensity and the proportion of the pulse wave / movement artifact in the collected signal data, and the signal data of the accelerometer reflects the information of the movement artifact, and the signal data of the sensor is the superposition of the movement artifact and the pulse wave signal, the larger the correlation coefficient, the greater the current exercise intensity and the greater the movement interference in the signal data of the sensor, and a larger step size factor needs to be set in the adaptive filter for denoising; on the contrary, if the correlation coefficient is smaller, the movement interference in the signal data of the sensor is smaller and the exercise intensity is lower, and a smaller step size factor can be used to obtain a more accurate pulse wave signal.

[0073] Step S24: Perform curve fitting on the M groups of second correlation coefficients and the second optimal step size factor to obtain a relational expression.

[0074] Among them, according to the relational expression obtained from the fitting curve of each correlation coefficient and each step size factor and the first correlation coefficient, the first optimal step size factor is obtained.

[0075] Preferably, according to the fitted curve, the relational expression is:

[0076] Or through linear fitting: Among them, is the correlation coefficient in the current motion scenario, μ * is the optimal step size factor in the current motion scenario, abs(·) and exp(·) are respectively the operations of taking the absolute value and the exponential function with the natural constant e as the base. See Figure 3 , which is a schematic diagram of curve fitting of a motion pulse wave denoising method provided by an embodiment of the present invention.

[0077] Preferably, the specific calculation formula of the correlation coefficient is the Pearson correlation coefficient calculation formula:

[0078]

[0079] Among them, pw i and acc i () are respectively the i-th signal data of the sensor and the i-th signal data on the j-th axis of the N-axis accelerometer, 1 ≤ j ≤ N; and are respectively the mean of the signal data of the sensor and the mean of the signal data on the j-th axis of the N-axis accelerometer; L is the length of extracting the sensor signal data and the N-axis accelerometer signal data; a(j) is the correlation coefficient between the sensor signal and the signal data on the j-th axis of the N-axis accelerometer; or,

[0080] Calculate using the Spearman correlation formula:

[0081]

[0082] Among them, is the rank difference between the i-th signal data of the sensor and the i-th signal data on the j-th axis of the N-axis accelerometer.

[0083] It should be noted that in addition to calculating using the above Pearson (Pearson) correlation coefficient and Spearman (Spearman) correlation coefficient formulas, the Kendall (Kendall) correlation coefficient can also be used for calculation. Among them, if the correlation coefficient is greater than 0, it indicates that the correlation between the sensor signal and the accelerometer signal is positively correlated. If the correlation coefficient is less than 0, it indicates that the correlation between the sensor signal and the accelerometer signal is negatively correlated. If the weak correlation coefficient is 1 or -1, it indicates that the sensor signal and the accelerometer signal can be described by a linear equation.

[0084] Step S103, filter the sensor signal according to the first optimal step size factor and a preset adaptive filtering algorithm to obtain a denoised pulse wave signal.

[0085] It should be noted that by converting the filtered pulse wave signal data into a spectrum and extracting the position with the highest frequency peak as the exercise heart rate frequency, the exercise heart rate value can be obtained.

[0086] The present invention calculates the correlation coefficient between the sensor signal and the N-axis accelerometer signal in different scenarios to describe the proportion of exercise intensity and pulse wave / movement artifact in the sensor signal. The larger the correlation coefficient, the greater the exercise intensity and the greater the movement interference in the sensor signal, and a larger step factor is required for denoising. It can more simply and conveniently reflect the value of the step factor. By curve fitting, a relational expression is obtained to accurately describe the relationship between the correlation coefficient and the step factor. Under different correlation coefficients, that is, in different exercise types and exercise intensities, the optimal step factor can be accurately found through the pre-fitted relational expression, thereby improving the denoising accuracy.

[0087] The present invention also provides an embodiment with a complete operation process. Refer to Figure 4 , which is a schematic flowchart of a method for denoising exercise pulse waves provided by an embodiment of the present invention, including steps S301 to S306, specifically as follows:

[0088] Step S301: Obtain the currently collected sensor signal as the input signal of the adaptive filter, and transmit the input signal to the adaptive filter. Exemplarily, select the LMS algorithm as the preset adaptive filter, and enter step S302.

[0089] Step S302: The adaptive filter processes the input signal and obtains an output signal, which can be expressed as:

[0090] y(n) = w T (n)x(n),

[0091] where w(n) is the weight vector of the adaptive linear combiner at this moment, x(n) is the input signal, and y(n) is the output signal. Then enter step S303.

[0092] Step S303: Obtain the desired signal, and enter step S304.

[0093] Step S304: Subtract the obtained desired signal from the calculated output signal to obtain an error signal, and the calculation is as follows:

[0094] e(n) = d(n) - y(n),

[0095] where e(n) is the error signal and d(n) is the desired signal. Then enter step S305.

[0096] Step S305: Update the weight vector of the adaptive linear combiner at the next moment according to the curve-fitted relational expression and the optimal step-size factor obtained in the current motion scenario, and the calculation is as follows:

[0097] w(n + 1)=w(n)+2μ * e(n)x(n),

[0098] where w(n + 1) is the weight vector of the adaptive linear combiner at the next moment, and μ * is the optimal step-size factor in the current motion scenario.

[0099] Then enter step S302, repeat steps S302 - S305 to reduce the mean square error, and finally enter step S306.

[0100] Step S306: Output the denoised pulse wave signal.

[0101] By using the LMS algorithm as the adaptive filter and according to the curve-fitted relational expression and the optimal step-size factor obtained in the current motion scenario, the present invention can update the weight vector of the adaptive linear combiner at the next moment. When the convergence condition is reached, the data of the denoised pulse wave signal is more accurate, thereby improving the accuracy and reliability of the obtained heart rate value.

[0102] Exemplarily, referring to Figure 5 , which is a comparison schematic diagram of a motion pulse wave before and after denoising provided by an embodiment of the present invention. The original pulse waveforms and triaxial accelerometer signals of the head / wrist / ankle and other parts under different exercise intensities are collected, and then several groups of sensor data with different exercise heart rates are selected from the collected data. The selected heart rate values include: 80 bpm, 95 bpm, 120 bpm, 140 bpm, 165 bpm. The data collection population is men aged 20 - 30 years old, and the sampling time for each group is 40 seconds. The sampling frequencies of the sensor and the accelerometer are both 50 Hz. Figure 5 a is the signal data of the original sensor, Figure 5 b is the pulse wave signal data after adaptive filtering denoising with the optimal step-size factor.

[0103] The present invention also provides a schematic diagram of the heart rate values after denoising for different exercise types and at different exercise intensities. Referring to Figure 6 , Figure 6 which is a schematic diagram of the heart rate values in different motion scenarios after denoising of a motion pulse wave provided by an embodiment of the present invention, Figure 6 a, Figure 6 b and Figure 6c represents the heart rate values during running, cycling, and basketball respectively. By detecting the heart rate of the head and chest, the heart rate values after denoising through adaptive filtering with the optimal step factor can adapt to different types and intensities of exercise, without the need to manually change the step factor each time.

[0104] See Figure 7 , which is a schematic structural diagram of a motion pulse wave denoising device provided by an embodiment of the present invention, including a first correlation coefficient calculation module 201, a first optimal step factor calculation module 202, and a pulse wave signal denoising module 203.

[0105] Among them, the first correlation coefficient calculation module 201 is used to obtain the currently collected sensor signal and the N-axis accelerometer signal, and calculate the correlation coefficient between the sensor signal and each axis accelerometer signal as the first correlation coefficient; where N is a positive integer.

[0106] The first optimal step factor calculation module 202 is used to obtain the optimal step factor as the first optimal step factor according to the relational expression of the curve fitting of each correlation coefficient and each step factor.

[0107] Specifically, collect the sensor signals and N-axis accelerometer signals in M different motion scenarios; where M is a positive integer.

[0108] Calculate the correlation coefficient of each group of sensor signals and N-axis accelerometer signals as the second correlation coefficient;

[0109] According to the preset adaptive filtering algorithm, adjust the step factors of M groups respectively, and select the step factor with the highest waveform amplitude and the most obvious period after filtering as the second optimal step factor in the current motion scenario, and obtain the corresponding M groups of step factors.

[0110] Specifically, extract the data of the sensor signal and the N-axis accelerometer data with the same time period and the same length, and calculate the correlation coefficient between the data of the sensor signal and each axis data of the N-axis accelerometer respectively, and obtain N groups of correlation coefficients as the second correlation coefficient.

[0111] Perform curve fitting on the M groups of second correlation coefficients and the second optimal step factor to obtain a relational expression.

[0112] Obtain the first optimal step factor according to the first correlation coefficient and the relational expression.

[0113] The pulse wave signal denoising module 203 is used to filter the sensor signal according to the first optimal step factor and the preset adaptive filtering algorithm to obtain the denoised pulse wave signal.

[0114] Preferably, the relational expression is specifically:

[0115] or wherein is the correlation coefficient in the current motion scenario, μ * is the optimal step size factor in the current motion scenario, abs(·) and exp(·) are the operations of taking the absolute value and the exponential function with the natural constant e as the base, respectively.

[0116] Preferably, the specific calculation formula of the correlation relationship is the Pearson correlation coefficient calculation formula:

[0117] where pw i and acc i (j) are the i-th signal data of the sensor and the i-th signal data on the j-th axis of the N-axis accelerometer, respectively, 1 ≤ j ≤ N; and are the mean values of the signal data of the sensor and the signal data on the j-th axis of the N-axis accelerometer, respectively; L is the length of the extracted sensor signal data and N-axis accelerometer signal data; a(j) is the correlation coefficient between the sensor signal and the signal data on the j-th axis of the N-axis accelerometer; or,

[0118] Calculate using the Spearman correlation formula:

[0119] wherein is the rank difference between the i-th signal data of the sensor and the i-th signal data on the j-th (1 ≤ j ≤ N) axis of the N-axis sensor.

[0120] By adopting the LMS algorithm as the adaptive filter and the relational expression obtained by curve fitting of each correlation coefficient and each correlation factor through the first optimal step size factor calculation module 202, the present invention can obtain the optimal correlation factor in the current motion scenario, make the denoising ability of the adaptive filter stronger, and thus obtain more accurate pulse wave signal data.

[0121] Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary general hardware. Of course, it can also be implemented by dedicated hardware including application specific integrated circuits, dedicated CPUs, dedicated memories, dedicated components, etc. Generally, functions completed by computer programs can be easily implemented by corresponding hardware, and the specific hardware structures for implementing the same function can also be various, such as analog circuits, digital circuits or dedicated circuits. However, for the present invention, software program implementation is a better embodiment in more cases. Based on such an understanding, the technical solution of the present invention, in essence or the part that makes contributions to the prior art, can be embodied in the form of a software product. The computer software product is stored in a readable storage medium, such as a floppy disk, USB flash drive, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disc of a computer, etc., and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments of the present invention.

[0122] The above is the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art of this technology, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements are also regarded as the protection scope of the present invention.

Claims

1. A method for denoising exercise pulse waves, characterized in that, the method includes: Obtain the currently collected sensor signals, N axis accelerometer signals, and calculate the correlation coefficients between the sensor signals and the axis accelerometer signals respectively as the first correlation coefficients; wherein, N is a positive integer; Obtaining an optimal step factor as the first optimal step factor according to the relational expression of curve fitting of each correlation coefficient and each step factor; Filtering the sensor signal according to the first optimal step factor and a preset adaptive filtering algorithm to obtain a denoised pulse wave signal; The specific relational expression is: μ * = 0.0043 * exp((abs( )) / 0.229) - 0.00279, or μ * = 0.14785 * abs( ) - 0.027, where is the correlation coefficient in the current motion scenario, μ * is the optimal step size factor in the current motion scenario, and are the operations of taking the absolute value and the exponential function with the natural constant e as the base, respectively; The obtaining of the optimal step factor as the first optimal step factor according to the fitting curve of each correlation coefficient and the step factor includes: Collection M sensor signals under different exercise scenarios and N axial accelerometer signals; where M is a positive integer Calculate the sum of each group of sensor signals N The correlation coefficient of the shaft accelerometer signal is the second correlation coefficient; According to the preset adaptive filtering algorithm, adjust respectively M the step size factors of the M groups, select the step size factor with the highest amplitude and the most obvious period of the filtered waveform as the second optimal step size factor in the current motion scenario, and obtain the corresponding M group of step size factors; Pair M Curve fitting is performed on the second correlation coefficient of the group and the second optimal step size factor to obtain a relational expression; Obtaining the first optimal step factor according to the first correlation coefficient and the relational expression.

2. The method for denoising exercise pulse waves according to claim 1, characterized in that, Calculating the correlation coefficient of each group of sensor signals and N The correlation coefficient of the shaft accelerometer signal is the second correlation coefficient, including: Extract the data of the sensor signals with the same time period and the same length, and N the data of the shaft accelerometer, and calculate the data of the sensor signals and N the correlation coefficient of each axis data of the shaft accelerometer respectively, and obtain N a set of correlation coefficients as the second correlation coefficient.

3. The method for denoising exercise pulse waves according to any one of claims 1-2, characterized in that, The specific calculation formula of the correlation coefficient is the Pearson correlation coefficient calculation formula: ， Among them, and are respectively the i th signal data of the sensor and N the j th signal data on the i axis of the axis accelerometer; and N are respectively the mean value of the sensor's signal data and j the mean value of the signal data on the L axis of the N axis accelerometer; is the length of the sensor signal data and N the j axis accelerometer signal data; or, Calculated using the Spearman correlation formula: ， Among them, is the i th signal data of the sensor and N the j th level difference of the signal data on the i th axis of the axis accelerometer.

4. An apparatus for denoising exercise pulse waves, characterized in that, it includes: The first correlation coefficient calculation module is used to obtain the currently collected sensor signal, N the axis accelerometer signal, and calculate the correlation coefficient between the sensor signal and each axis accelerometer signal respectively as the first correlation coefficient; where N is a positive integer; A first optimal step factor calculation module, configured to obtain an optimal step factor as the first optimal step factor according to the relational expression of curve fitting of each correlation coefficient and each step factor; A pulse wave signal denoising module, configured to filter the sensor signal according to the first optimal step factor and a preset adaptive filtering algorithm to obtain a denoised pulse wave signal; The specific relational expression is: μ * = 0.0043 * exp((abs( )) / 0.229) - 0.00279, or μ * = 0.14785 * abs( ) - 0.027, where is the correlation coefficient in the current motion scenario, μ * is the optimal step size factor in the current motion scenario, and are the operations of taking the absolute value and the exponential function with the natural constant e as the base, respectively; The first optimal step factor calculation module is specifically: Collection M Sensor signals and N axial accelerometer signals under different exercise scenarios; where M is a positive integer; Calculate the sum of each group of sensor signals N The correlation coefficient of the shaft accelerometer signal is the second correlation coefficient; According to the preset adaptive filtering algorithm, adjust respectively M the step size factors of the M group, select the step size factor with the highest amplitude and the most obvious period of the waveform after filtering as the second optimal step size factor under the current motion scenario, and obtain the corresponding M group of step size factors; Pair M Curve fitting is performed on the second correlation coefficient and the second optimal step factor of the group to obtain a relational expression; Obtaining the first optimal step factor according to the first correlation coefficient and the relational expression.

5. The apparatus for denoising exercise pulse waves according to claim 4, characterized in that, Calculating the correlation coefficient of each group of sensor signals and N The correlation coefficient of the axis accelerometer signal is the second correlation coefficient, including: Extract the data of the sensor signals with the same time period and the same length and N axis accelerometer data, and calculate the data of the sensor signals and N the correlation coefficient of each axis data of the axis accelerometer respectively, and obtain N a set of correlation coefficients as the second correlation coefficient.

6. The apparatus for denoising exercise pulse waves according to any one of claims 4-5, characterized in that, The specific calculation formula of the correlation coefficient is the Pearson correlation coefficient calculation formula: ， Among them, and are respectively the i th signal data of the sensor and N the j th signal data on the i axis of the accelerometer; and are respectively the mean value of the sensor's signal data and N the mean value of the signal data on the j axis of the accelerometer; L is the length of extracting the sensor signal data and N the signal data of the axis accelerometer; N is the correlation coefficient of the sensor signal and j the signal data on the axis of the accelerometer; or, Calculated using the Spearman correlation formula: ， Among them, is the i th signal data of the sensor and N the j th i level difference of the signal data on the

Images