CN117110649A

CN117110649A - Quality enhancement method, device and system for motion data

Info

Publication number: CN117110649A
Application number: CN202310969242.0A
Authority: CN
Inventors: 王耘田; 任俊宏; 张文生
Original assignee: Institute of Automation of Chinese Academy of Science
Current assignee: Institute of Automation of Chinese Academy of Science
Priority date: 2023-08-02
Filing date: 2023-08-02
Publication date: 2023-11-24

Abstract

The invention provides a quality enhancement method, a device and a system for motion data, wherein the method comprises the following steps: acquiring motion data, wherein the motion data are acceleration data obtained by data acquisition when three axes in a triaxial acceleration sensor point to the gravity direction respectively; determining three axial calibration functions, and determining three axial calibration values based on the calibration functions and the motion data; based on the calibration value and the gravity acceleration in the gravity direction, performing parameter optimization on the calibration function to obtain an optimal calibration function; based on the optimal calibration functions in the three axial directions, the acceleration data in the three axial directions in the motion data are subjected to offset calibration to obtain calibration motion data, the defects of large error and low quality of the acquired motion data caused by hardware difference in the traditional scheme are overcome, the sampling error is reduced through offset calibration of the motion data, the data quality is improved, and the quality problem of the motion data can be effectively improved on the premise of limited hardware level.

Description

Quality enhancement method, device and system for motion data

Technical Field

The present invention relates to the field of data processing technologies, and in particular, to a method, an apparatus, and a system for enhancing quality of motion data.

Background

The triaxial acceleration sensor is one of acceleration sensors, works based on the basic principle of acceleration, has the characteristics of small volume, light weight and the like, can measure the spatial acceleration, reflects the motion property of an object, and is the preferred equipment for collecting the motion data of a human body. At present, when the motion data of a human body is collected, the motion data is usually required to be monitored for a long time so as to better collect and store the data, thereby the motion states and the characteristics under different physical activity intensities are obtained through the analysis of acceleration data in three axial directions in the motion data.

However, because of uneven hardware processing level of the triaxial acceleration sensor and differences in design of the motion monitoring product of the built-in accelerometer, errors and noise can occur in the monitored motion data when the motion monitoring is performed according to the design, so that the quality of the data is affected. In addition, if the motion monitoring period is long, the sensor needs to continuously acquire data for a long time, which can continuously amplify the influence of hardware problems on the data quality, so that the error of the motion data finally acquired is large and the data quality is low. Therefore, how to improve the quality of motion data under the condition that the hardware level is limited becomes a current problem to be solved.

Disclosure of Invention

The invention provides a quality enhancement method, device and system of motion data, which are used for solving the defect that the influence of hardware difference on the quality of acquired motion data in the prior art, and carrying out offset calibration on the motion data through gravity calibration so as to enable the motion data to be closer to actual data, reduce data sampling errors caused by hardware reasons and improve the data quality of the motion data.

The invention provides a quality enhancement method of motion data, which comprises the following steps:

acquiring motion data, wherein the motion data are acceleration data obtained by data acquisition when three axes in a triaxial acceleration sensor point to the gravity direction respectively;

determining calibration functions in the three axial directions, and determining calibration values in the three axial directions based on the calibration functions and the motion data;

based on the calibration value and the gravity acceleration in the gravity direction, performing parameter optimization on the calibration function to obtain an optimal calibration function;

and carrying out offset calibration on the acceleration data in the three axial directions in the motion data based on the optimal calibration functions in the three axial directions to obtain calibration motion data.

According to the quality enhancement method of the motion data, the calibration function comprises offset and calibration sensitivity;

-said determining calibration values in said three axial directions based on said calibration function and said motion data; based on the calibration value and the gravitational acceleration in the gravitational direction, performing parameter optimization on the calibration function to obtain an optimal calibration function, including:

determining calibration values in the three axial directions based on the motion data and initial values of the calibration function and its bias and calibration sensitivity;

determining an optimal value of bias and calibration sensitivity in the three-axis calibration function based on the calibration value and the gravitational acceleration;

and carrying out parameter optimization on the offset and the calibration sensitivity in the three axial calibration functions based on the optimal values to obtain an optimal calibration function.

According to the method for enhancing the quality of the motion data provided by the invention, the method for determining the optimal values of offset and calibration sensitivity in the three axial calibration functions based on the calibration values and the gravitational acceleration comprises the following steps:

residual calculation is carried out based on the calibration value and the gravity acceleration, so as to obtain residual values in the three axial directions;

and carrying out parameter optimization through a least square method based on the residual value to obtain the optimal values of offset and calibration sensitivity in the calibration functions in the three axial directions.

According to the method for enhancing the quality of the motion data provided by the invention, the offset calibration is performed on the acceleration data in the three axial directions in the motion data based on the optimal calibration function in the three axial directions, so as to obtain the calibrated motion data, and then the method further comprises the following steps:

determining a first array corresponding to each of the three axial acceleration data based on the three axial acceleration data in the calibration motion data;

equal-length segmentation is carried out on the acceleration data in the first array to obtain a plurality of subarrays;

based on the range of each subarray, carrying out filtering processing on the acceleration data in the first array to obtain filtered acceleration data;

and determining filtered motion data based on the filtered acceleration data in the three axes and the acceleration data in the three axes in the calibration motion data.

According to the quality enhancement method of motion data provided by the invention, based on the extreme differences of all the subarrays, the acceleration data in the first array is subjected to filtering processing to obtain filtered acceleration data, and the method comprises the following steps:

determining target acceleration data corresponding to the acceleration data in any one of the subarrays from the acceleration data in the second array under the condition that the range of any one of the subarrays is smaller than or equal to a difference threshold value, wherein the acceleration data in the second array is obtained by filtering based on the acceleration data in the first array;

Replacing the acceleration data in any subarray with the target acceleration data to obtain the filtered acceleration data corresponding to any subarray;

under the condition that the range of any subarray is larger than a difference value threshold value, determining the acceleration data in any subarray as the filtered acceleration data corresponding to any subarray;

and determining the corresponding axial filtered acceleration data based on the filtered acceleration data corresponding to each subarray.

According to the method for enhancing the quality of the motion data provided by the invention, the step of determining the filtered motion data based on the filtered acceleration data in the three axial directions and the acceleration data in the three axial directions in the calibration motion data comprises the following steps:

determining acceleration data in a vector and direction based on acceleration data in the three axial directions in the calibration motion data;

filtered motion data is determined based on the acceleration data in the vector sum direction and the filtered acceleration data in the three axes.

The invention also provides a quality enhancement device of motion data, comprising:

the acquisition unit is used for acquiring motion data, wherein the motion data are acceleration data obtained by data acquisition when three axes in the triaxial acceleration sensor point to the gravity direction respectively;

A calculation unit configured to determine calibration functions in the three axial directions, and determine calibration values in the three axial directions based on the calibration functions and the motion data;

the optimization unit is used for carrying out parameter optimization on the calibration function based on the calibration value and the gravity acceleration in the gravity direction to obtain an optimal calibration function;

and the calibration unit is used for carrying out offset calibration on the acceleration data in the three axial directions in the motion data based on the optimal calibration functions in the three axial directions to obtain calibration motion data.

The invention also provides a quality enhancement system of the motion data, which comprises a triaxial acceleration sensor and a quality enhancement device of the motion data;

the three-axis acceleration sensor is used for acquiring data when three axes point to the gravity direction respectively to obtain motion data, and transmitting the motion data to the quality enhancement device of the motion data;

the quality enhancement device of the motion data is used for acquiring the motion data, determining the calibration values of the three axial directions based on the motion data and the calibration functions of the three axial directions, performing parameter optimization on the calibration functions based on the calibration values and the gravitational acceleration of the gravitational direction to obtain an optimal calibration function, and performing offset calibration on the acceleration data of the three axial directions in the motion data based on the optimal calibration function of the three axial directions to obtain calibration motion data.

The invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing a method of enhancing the quality of motion data as described in any of the above when executing the program.

The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method of quality enhancement of motion data as described in any of the above.

According to the quality enhancement method, device and system for the motion data, the three axial calibration values are determined through the three axial calibration functions and the motion data, and the motion data are acceleration data obtained by data acquisition when three axes in the triaxial acceleration sensor point to the gravity direction respectively; according to the calibration value and the gravity acceleration in the gravity direction, parameter optimization is carried out on the calibration function, bias calibration is carried out on acceleration data in three axial directions in the motion data according to the optimized optimal calibration function, so that the calibration motion data are obtained, the quality improvement of the motion data is realized, the defects of large error and low quality of the acquired motion data caused by hardware difference in the traditional scheme are overcome, the sampling error is reduced through the bias calibration of the motion data, the data quality is improved, the quality problem of the motion data can be effectively improved on the premise of limited hardware level, and a foundation is laid for smooth development of motion monitoring and improvement of the motion monitoring accuracy.

Drawings

In order to more clearly illustrate the invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.

FIG. 1 is a flow chart of a method for enhancing the quality of motion data provided by the invention;

FIG. 2 is a graph comparing the effects of motion data provided by the present invention before and after offset calibration;

FIG. 3 is a graph comparing the effects of calibration motion data provided by the present invention before and after filtering;

FIG. 4 is a general flow chart of a bias calibration process for motion data provided by the present invention;

FIG. 5 is a general flow chart of a process for filtering calibration motion data provided by the present invention;

FIG. 6 is a schematic diagram of a quality enhancement device for motion data provided by the present invention;

FIG. 7 is a schematic diagram of a quality enhancement system for motion data provided by the present invention;

fig. 8 is a schematic structural diagram of an electronic device provided by the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

At present, when motion monitoring is carried out through a triaxial acceleration sensor, the acquired motion data can generate errors and contain noise due to uneven hardware processing level of the triaxial acceleration sensor and differences in product design, and the data quality is reduced.

Further, if the motion monitoring period is long, the sensor needs to continuously collect data for a long time, which can make the influence of hardware problems on the data quality be amplified continuously, so that the error of the motion data finally collected is large and the data quality is low. For this situation, attention is paid to the hardware design level and the processing technology level, that is, the difference of hardware is improved through optimization of the hardware design of the product, improvement of the processing technology and the like, so that the quality of the acquired motion data is improved. However, improvements to hardware design and processing techniques often require significant resources, are too costly, are relatively long-lived, and are not practical.

In this regard, the invention provides a quality enhancement method of motion data, which aims to improve from the software algorithm level, and reduces sampling errors by carrying out offset calibration on the motion data, so that the quality of the data is improved, and further, the quality problem of the motion data can be effectively improved on the premise of limited hardware level, so as to lay a foundation for smoothly developing motion monitoring and breaking through the efficiency and accuracy of the motion monitoring. FIG. 1 is a schematic flow chart of a method for enhancing quality of motion data according to the present invention, as shown in FIG. 1, applied to a triaxial acceleration sensor, the method comprising:

Step 110, acquiring motion data, wherein the motion data is acceleration data obtained by data acquisition when three axes in a triaxial acceleration sensor point to the gravity direction respectively;

specifically, before the quality enhancement, firstly, the motion data to be processed, namely, the motion data with errors and the motion data with error correction and quality recovery are required to be obtained, and because the triaxial acceleration sensor has the advantages of small volume, light weight, portability and the like, can measure the spatial acceleration and can reflect the motion property, the triaxial acceleration sensor is selected as the data acquisition equipment for monitoring and acquiring the motion data in the embodiment of the invention.

Here, the triaxial acceleration sensor may be of a piezoresistive type, a piezoelectric type or a capacitive type, and embodiments of the present invention are not limited thereto in particular. The acceleration generated during data acquisition based on the acceleration is proportional to the changes of resistance, voltage and capacitance, and the data can be acquired through corresponding amplifying and filtering circuits. In addition, it is worth noting that in the embodiment of the invention, the motion monitoring based on the triaxial acceleration sensor comprises the monitoring of physical activity and the monitoring of static activity, and the motion states and characteristics under different physical activity intensities can be obtained through analyzing the acceleration data in three axial directions in the motion data collected by the monitoring.

It may be understood that in the process of actually performing motion monitoring and collecting motion data, the triaxial acceleration sensor may be started first, and the sampling frequency may be configured for the triaxial acceleration sensor so as to perform data collection, thereby obtaining motion data, where the triaxial acceleration sensor used for data collection may be the triaxial acceleration sensor itself, or a data collection device with the triaxial acceleration sensor built therein, or a data collection system connected with the triaxial acceleration sensor.

Specifically, after the triaxial acceleration sensor is started, three axes x-axis, y-axis and z-axis of the triaxial acceleration sensor may be respectively allowed to stand in a gravity direction (g-direction), data of a period of time may be collected, acceleration data in the axial direction (gravity axial direction) may be taken, and motion data may be obtained, where the motion data may be expressed as x= [ x1, x2, …, xn ], y= [ y1, y2, …, yn ], z= [ z1, z2, …, zn ]. For example, a sampling frequency of 30HZ may be configured to collect acceleration data for about 5 minutes, and after the collection is completed, an offline data set containing about 10000 samples may be obtained.

Here, taking the case that the z-axis points in the gravity direction as an example, when the z-axis points in the gravity direction, the data set includes x= [ x1, x2, …, x10000], y= [ y1, y2, …, y10000], z= [ z1, z2, …, z10000]. And because the offset calibration method for correcting the error of the motion data adopts gravity calibration, only the acceleration data z collected in the gravity direction is adopted. Similarly, when the x axis and the y axis point to the gravity direction respectively, the acceleration data x and y collected in the gravity direction are selected as well, and finally the motion data to be processed can be obtained.

Ideally, 10000 sample values contained in acceleration data z acquired in the gravitational direction should be equal to the gravitational acceleration g=9.8 m/s ² Equal, i.e. z= [9.8,9.8, …,9.8]However, the actual sample is not 9.8, i.e. the actual observed value of the sample in z is often not 9.8, because the sample is error when the device is stationary during the actual sampleThere is an error/bias. Therefore, error correction is required to improve the data quality, i.e. steps 120 to 140 are performed to reduce sampling errors due to hardware reasons and restore the data quality.

Here, it is worth noting that, unlike the actual data acquisition performed by the triaxial acceleration sensor, the obtained actual human motion data is, in the embodiment of the present invention, in order to correct the sampling error when the device is stationary, the axial direction is pointed to the gravity direction in a targeted manner, and the acceleration data in the gravity direction is acquired for performing offset calibration, and in the process of actually acquiring the human motion data, the acceleration data in the gravity direction does not need to be acquired specially. In short, the motion data in the embodiment of the invention is fixedly acquired corresponding to the offset calibration process of the gravity calibration, and once the optimal calibration function is determined, the data acquisition can be normally performed to obtain the real human motion data without specially acquiring the acceleration data in the gravity direction.

Step 120, determining calibration functions in three axial directions, and determining calibration values in three axial directions based on the calibration functions and the motion data;

step 130, performing parameter optimization on the calibration function based on the calibration value and the gravity acceleration in the gravity direction to obtain an optimal calibration function;

and 140, performing offset calibration on acceleration data in three axial directions in the motion data based on the optimal calibration functions in the three axial directions to obtain calibration motion data.

Specifically, after obtaining the motion data through step 110, the calculation of the calibration values in three axial directions can be performed according to the motion data, the calibration function is optimized according to the calibration values, and the offset calibration is performed according to the optimized calibration function, which specifically may include:

firstly, determining three axial calibration functions, namely establishing the calibration functions for acceleration data in the three axial directions, wherein the calibration functions comprise parameter offset and calibration sensitivity, and setting initial values for the offset and the calibration sensitivity before calculating according to the calibration functions;

then, the calibration function and the motion data can be used for calculating the calibration value in each axial direction, namely the motion data can be substituted into the calibration function for solving so as to obtain the calibrated values of the acceleration data in the three axial directions, namely the calibration values in the three axial directions.

The calibration value calculated here is based on the initial values of the set offset and the calibration sensitivity, and the set initial values often do not correspond to the optimal values in the scene and the axial direction, so the calculated calibration value cannot be used as the final value at this time, and the calibration function needs to be optimized, so that the final calibration motion data is obtained according to the optimized calibration function.

I.e. performing optimization calculation according to the calibration value to update and optimize parameters of the calibration function, thereby obtaining an optimized calibration function, which may be referred to herein as an optimal calibration function, specifically, on the basis of the calibration value in the corresponding axial direction, combining the gravitational acceleration in the gravitational direction (g=9.8 m/s ² ) And performing optimization calculation to perform parameter optimization on the calibration function, thereby obtaining an optimized calibration function, namely an optimal calibration function in three axial directions.

And then, performing offset calibration on the motion data according to the optimal calibration function to obtain final calibration data, namely calibration motion data, wherein specifically, offset calibration can be performed on acceleration data in three axial directions in the motion data according to the optimal calibration function in three axial directions respectively, namely, the acceleration data in the x axial direction in the motion data is calibrated by utilizing the optimal calibration function in the x axial direction, and errors/offsets in the acceleration data are calibrated, so that the calibrated acceleration data in the x axial direction, namely the calibrated motion data in the x axial direction, is obtained.

Similarly, the optimal calibration function in the y-axis can be utilized to calibrate the acceleration data in the y-axis in the motion data, the optimal calibration function in the z-axis is utilized to calibrate the acceleration data in the z-axis in the motion data, the calibration motion data in the y-axis and the acceleration data in the z-axis are obtained, and the calibration motion data can be obtained after summarizing.

FIG. 2 is a graph showing the results of the motion data before and after offset calibration, as shown in FIG. 2, comparing the acceleration data before (before) and after (after) calibration, it can be seen that the offset calibration by gravity calibration can correct the offset caused by the hardware difference in the data acquisition process, so that the acceleration data in the acquired gravity direction can approach the gravity acceleration g=9.8m/s after the calibration (after) ² Even equal to g, sampling errors are reduced, thereby improving the data quality of the motion data.

The method is different from the traditional scheme that the quality improvement of motion data starts from the hardware design and processing technology level, and the method is used for processing from the software algorithm level, so that the motion data is closer to real data through offset calibration, sampling errors caused by hardware are reduced, the data quality is improved, and meanwhile the problems of high cost, long period, high difficulty and poor feasibility existing in the traditional scheme that the data quality is improved through hardware improvement are solved.

According to the embodiment of the invention, the motion data is processed by the gravity calibration offset calibration method, so that the quality of the motion data is improved rapidly and accurately by simple operation under the condition of not depending on additional equipment such as a turntable and strict experimental conditions, the quality of the motion data is improved effectively, the improvement of the data quality is realized, and a powerful support is provided for the smooth development of motion monitoring and the breakthrough of the motion monitoring efficiency and accuracy.

According to the quality enhancement method of the motion data, the three axial calibration values are determined through the three axial calibration functions and the motion data, and the motion data are acceleration data obtained by data acquisition when the three axes in the triaxial acceleration sensor point to the gravity direction respectively; according to the calibration value and the gravity acceleration in the gravity direction, parameter optimization is carried out on the calibration function, bias calibration is carried out on acceleration data in three axial directions in the motion data according to the optimized optimal calibration function, so that the calibration motion data are obtained, the quality improvement of the motion data is realized, the defects of large error and low quality of the acquired motion data caused by hardware difference in the traditional scheme are overcome, the sampling error is reduced through the bias calibration of the motion data, the data quality is improved, the quality problem of the motion data can be effectively improved on the premise of limited hardware level, and a foundation is laid for smooth development of motion monitoring and improvement of the motion monitoring accuracy.

Based on the above embodiment, the calibration function includes the bias and the calibration sensitivity;

determining calibration values in three axial directions based on the calibration function and the motion data; based on the calibration value and the gravitational acceleration in the gravitational direction, performing parameter optimization on the calibration function to obtain an optimal calibration function, including:

determining calibration values in three axial directions based on the motion data and initial values of the calibration function and its offset and calibration sensitivity;

determining an optimal value of bias and calibration sensitivity in a calibration function in three axial directions based on the calibration value and the gravitational acceleration;

and carrying out parameter optimization on the offset and the calibration sensitivity in the calibration functions in the three axial directions based on the optimal values to obtain the optimal calibration function.

Specifically, in step 120 and step 130, the process of determining calibration values in three axial directions according to the calibration function and the motion data, and optimizing parameters of the calibration function according to the calibration values and the gravitational acceleration in the gravitational direction to obtain an optimal calibration function may specifically include:

firstly, when the calibration functions in the three axial directions are established, initial values of offset and calibration sensitivity are respectively set for the calibration functions in the three axial directions, so that the initial values of offset and calibration sensitivity can be assigned to corresponding parameters in the calibration functions in the corresponding axial directions to form a complete calibration equation only comprising input and output unknown numbers, namely, the three complete calibration functions in the axial directions; the calibration function can be expressed herein as:

acc_cali＝(acc_raw+offset)*r

Wherein acc_raw is input, that is, acceleration data in the corresponding axial direction in the motion data, acc_cali is output, that is, a calibration value in the corresponding axial direction, offset is offset, and r represents the calibration sensitivity.

The motion data may then be substituted into this complete calibration equation/function to obtain an output value of the calibration function, i.e. a calibration value, where in particular acceleration data in three axes of the motion data are substituted into the calibration equation/function in the corresponding axes, respectively, to solve for the output value, i.e. the calibration value in the three axes.

Subsequently, a calculation can be performed based on the calibration value and the gravitational acceleration in the gravitational direction to solve for the optimum value of the bias and the calibration sensitivity in the calibration function, which here may be specifically performed by comparing the calibration value in the corresponding axial direction and the gravitational acceleration in the gravitational direction (g=9.8 m/s ² ) And combining to form a data pair, and performing optimization calculation according to the data pair, so that the optimal values of the offset and the calibration sensitivity in the calibration functions in the corresponding axial directions can be obtained, and the calculation process is repeated, so that the optimal values of the offset and the calibration sensitivity in the calibration functions in the three axial directions can be finally obtained.

And then, updating and optimizing the optimal value to obtain an optimal calibration function, namely, parameter optimization can be carried out on the calibration function in the corresponding axial direction according to the optimal value of the offset and the calibration sensitivity in each axial direction, and the value of the offset and the calibration sensitivity in the updated function can be specifically replaced by the optimal value, so that the updated calibration function, namely, the optimal calibration function, is obtained. The optimal calibration function can be expressed here as:

acc_cali _x ＝(acc_raw _x +offset _x )*r _x

acc_cali _y ＝(acc_raw _y +offset _y )*r _y

acc_cali _z ＝(acc_raw _z +offset _z )*r _z

In the acc_raw _x For input of a calibration function in the x-axis, i.e. acceleration data in the x-axis of the motion data, acc_cali _x As a calibration function in the x-axis directionOutput of (a), i.e. calibration value in x-axis, offset _x Is the optimum value of offset in the calibration function in the x-axis direction, r _x Is the optimal value of the calibration sensitivity in the calibration function in the x-axis.

acc_raw _y For input of a calibration function in the y-axis, i.e. acceleration data in the y-axis of the motion data, acc_cali _y For output of the calibration function in the y-axis, i.e. the calibration value in the y-axis, offset _y Is the optimum value of the offset in the calibration function in the y-axis direction, r _y Is the optimal value of the calibration sensitivity in the calibration function in the y-axis.

acc_rw _z For input of a calibration function in the z-axis, i.e. acceleration data in the z-axis in the motion data, cc_cali _z For output of calibration function in z-axis, i.e. calibration value in z-axis, offset _z Is the optimal value of the offset in the calibration function in the z-axis direction, r _z Is the optimal value of the calibration sensitivity in the calibration function in the z-axis.

Based on the above embodiment, determining the optimum values of bias and calibration sensitivity in the calibration function in three axial directions based on the calibration values, and the gravitational acceleration, includes:

Residual calculation is carried out based on the calibration value and the gravity acceleration, so as to obtain residual values in three axial directions;

and (3) carrying out parameter optimization by a least square method based on the residual value to obtain the optimal values of offset and calibration sensitivity in the calibration functions in three axial directions.

Specifically, the process of determining the optimum values of the offset and the calibration sensitivity in the calibration functions in the three axial directions according to the calibration values and the gravitational acceleration may specifically include:

first, residual calculation can be performed based on the calibration value and the gravitational acceleration to obtain residual values in three axial directions, that is, the residual values can be calculated based on the calibration value and the gravitational acceleration in the gravitational direction (g=9.8 m/s ² ) Calculating residual errors to obtain residual values, specifically, combining the calibration values and the gravitational acceleration to form data pairs, obtaining three data pairs in the axial direction, and then obtaining the data pairs in each axial directionThe data pairs are all subjected to residual calculation to obtain residual values in each axial direction, and the formula of the residual calculation can be as follows:

Res＝g-acc_cali

wherein Res represents the residual value, g is the gravitational acceleration, g=9.8 m/s ² acc_cali is the calibration value in the corresponding axial direction.

And then, carrying out optimization calculation according to the residual value to obtain the optimal value of the offset and the calibration sensitivity in the calibration function in the corresponding axial direction, wherein the parameter optimization can be carried out by a least square method on the basis of the residual value, so as to obtain the optimal values of the offset and the calibration sensitivity in the calibration function in the three axial directions. Here, the calculation formula of the optimal value can be expressed as:

Where (offset, r) is the optimum value of the offset and calibration sensitivity in the corresponding calibration function in the axial direction, n is the number of samples, i.e. the number of acceleration data in the corresponding axial direction in the motion data, and Res is the residual value in the corresponding axial direction.

Based on the above embodiment, based on the optimal calibration functions in the three axial directions, offset calibration is performed on acceleration data in the three axial directions in the motion data, so as to obtain calibrated motion data, and then the method further includes:

determining a first array corresponding to each of the three axial acceleration data based on the acceleration data in the three axial directions in the calibration motion data;

the filtered motion data is determined based on the filtered acceleration data in the three axes and the acceleration data in the three axes in the calibration motion data.

Considering that the motion data obtained by data acquisition through the triaxial acceleration sensor has the problem of sampling errors and noise, namely certain noise components are often contained in the motion data, after the offset calibration of the motion data is completed through the offset calibration method of gravity calibration, the calibration motion data can be subjected to filtering processing (self-adaptive filtering processing) so as to filter the noise components in the motion data, and the data quality is further optimized.

Specifically, after the motion data is offset calibrated to obtain calibration motion data, in the process of performing filtering processing, the calibration motion data may be first processed to generate a corresponding array, that is, acceleration data in three axial directions are extracted respectively to form a corresponding array, so that an array s_raw corresponding to the acceleration data in three axial directions may be obtained, and for convenience of distinguishing, the array s_raw is referred to herein as a first array.

Then, the acceleration data in the first array s_raw may be equally divided to obtain a plurality of subarrays, i.e. the acceleration data in the s_raw may be grouped according to equal-length slices to obtain a plurality of subarrays, where each subarray S includes n samples, s= [ S1, S2, …, sn ]. Here, the value of the length n affects the final result, so in the embodiment of the present invention, the sampling frequency may be preferably used as the reference value of the packet length n.

Then, the range of each sub-array can be calculated, that is, the difference Diff (S) =max (S) -Min (S) between the maximum value and the minimum value in the sub-array is calculated, and then the acceleration data in the first array can be subjected to filtering processing according to the difference, so as to obtain filtered acceleration data, which may specifically be that when the range of each sub-array is determined to need to be subjected to filtering processing, the acceleration data in the first array is subjected to filtering processing, so as to obtain filtered acceleration data, which is referred to as filtered acceleration data herein.

Specifically, whether the acceleration data in the corresponding axial direction in the calibration motion data needs to be filtered or not can be judged according to the range of the subarray, specifically, the acceleration data in the subarray in the corresponding axial direction in the acceleration data needs to be filtered, that is, when the range reflects that the difference between the acceleration data in the corresponding subarray is smaller (for example, 0.1), the fact that the acceleration data in the subarray contains noise can cause motion data distortion is indicated, and therefore the acceleration data in the subarray needs to be filtered and noise reduced; on the contrary, when the extremely poor reflection shows that the difference between the acceleration data in the corresponding subarray is larger, the acceleration data in the subarray does not contain noise, and the motion data is close to the true value, so that the filtering processing is not needed.

Further, when the filtering processing is required to be performed in the extremely poor reflection, the filtering processing can be performed on the acceleration data in the corresponding subarray so as to obtain the filtered acceleration data. Correspondingly, when the extremely poor reflection is not needed to be subjected to filtering processing, the acceleration data in the corresponding subarray can be directly reserved and used as the filtering acceleration data corresponding to the corresponding subarray. And traversing all the subarrays and all the axial directions to obtain the filtered acceleration data corresponding to the acceleration data in the calibration motion data.

Then, final filtering data, namely final filtering motion data obtained by filtering the calibration motion data, can be determined according to the three axial filtering acceleration data; here, specifically, on the basis of the filtered acceleration data in the three axial directions, calculation may be performed in combination with the acceleration data in the three axial directions in the calibration motion data to obtain the filtered motion data.

Based on the above embodiment, based on the range of each subarray, the filtering processing is performed on the acceleration data in the first array to obtain filtered acceleration data, including:

determining target acceleration data corresponding to the acceleration data in any one of the subarrays from the acceleration data in the second subarray under the condition that the range of the subarray is smaller than or equal to a difference threshold value;

replacing the acceleration data in the subarray with the target acceleration data to obtain the filtered acceleration data corresponding to the subarray, wherein the acceleration data in the second array is obtained by filtering based on the acceleration data in the first array;

under the condition that the range of any subarray is larger than a difference value threshold value, determining the acceleration data in the subarray as the filtered acceleration data corresponding to the subarray;

Specifically, the process of filtering the acceleration data in the first array according to the range of each sub array to obtain filtered acceleration data specifically includes the following steps:

first, a threshold value corresponding to the range, i.e., a difference threshold value, may be determined, and this difference threshold value is used to characterize a critical value of the filtering process, i.e., a minimum value of the range (difference between maximum value and minimum value) between acceleration data caused by tolerable/acceptable noise.

Then, a status function can be established according to the difference threshold to determine whether the acceleration data in the corresponding subarray needs to be filtered, where the status function can representWherein the initial value of state (S) is 0 and a is the difference threshold.

The difference threshold can be obtained by sinusoidal vibration test, the sinusoidal vibration test can be performed under different frequency and amplitude conditions to test the acquisition accuracy of the triaxial acceleration sensor, the experimental data is visualized, a section of data containing more noise is selected from the data to be sliced to obtain m groups of data, the difference Diff between the maximum value and the minimum value in each group of data is calculated respectively, and the average value of the Diff in all groups is obtained, namely This average value is taken as a reference value for the difference threshold a. In addition, a section of real motion data containing different motion amplitudes can be acquired, more noise is observed when the noise is generated, and the data of the corresponding section is intercepted to calculate the difference threshold value a.

Further, when the value of the state function is 0, that is, when the range of the corresponding subarray is smaller than or equal to the difference threshold, filtering processing is required at this time, specifically, the acceleration data in the corresponding subarray in the first array is processed by using the acceleration data in the second array after filtering processing, so as to obtain the filtered acceleration data corresponding to the corresponding subarray. The acceleration data in the second array s_filter is obtained by performing kalman filtering fk on the acceleration data in the first array s_raw.

That is, when the range of the corresponding subarray is less than or equal to the difference threshold, the acceleration data corresponding to the acceleration data in the subarray can be intercepted from the second array as target acceleration data; that is, when state (S) =0, acceleration data (kalman-filtered acceleration data) fk (S) corresponding to S is intercepted from s_filter as target acceleration data. And then, the target acceleration data can replace the acceleration data in the corresponding subarray, so that the filtered acceleration data corresponding to the subarray is obtained, namely, the target acceleration data fk (S) is used for replacing the acceleration data in the subarray S, and the filtered acceleration data corresponding to the S is obtained.

Correspondingly, when the value of the state function is 1, that is, when the range of the corresponding subarray is greater than the difference threshold, filtering processing is not needed at the moment, so that the acceleration data in the subarray can be directly used as the filtering acceleration data corresponding to the subarray; that is, when state (S) =1, the acceleration data in S remains unchanged as the filtered acceleration data corresponding to S. And repeating the process, and traversing all the subarrays to obtain the filtered acceleration data corresponding to each subarray.

And finally, determining the corresponding axial filtering acceleration data according to the filtering acceleration data corresponding to each sub-array, namely determining a plurality of sub-arrays which are segmented by the first array corresponding to the acceleration data in each axial direction by taking the axial direction as a unit, determining the axial filtering acceleration data according to the filtering acceleration data corresponding to the plurality of sub-arrays, and traversing all axial directions to obtain three axial filtering motion data, wherein the three axial filtering motion data is the final motion data obtained by adaptively filtering the calibration motion data to remove noise components.

Fig. 3 is a comparison chart of the effects of the calibration motion data before and after filtering, as shown in fig. 3, the acceleration data after the comparison of the adaptive filtering (raw) can be used for removing noise generated under different motion conditions due to hardware difference in the sampling process by the adaptive filtering process, so that the data quality is further improved.

Based on the above embodiment, determining the filtered motion data based on the filtered acceleration data in three axes and the acceleration data in three axes in the calibration motion data includes:

determining acceleration data in a vector sum direction based on acceleration data in three axes in the calibration motion data;

the filtered motion data is determined based on the acceleration data in the vector sum direction and the filtered acceleration data in the three axes.

Specifically, the process of determining the filtered motion data according to the filtered acceleration data in three axial directions and the acceleration data in three axial directions in the calibration motion data includes:

firstly, the acceleration data in the three axial directions in the calibration motion data can be calculated, and the acceleration data in the three axial directions in the calibration motion data can be utilized, and the acceleration data in the vector directions can be obtained through calculation by using a vector summation formula, wherein the calculation formula can be expressed asWherein x, y and z are acceleration data in the x-axis, y-axis and z-axis respectively, and ENMO is acceleration data in the vector and direction;

And then, determining the filtered motion data according to the acceleration data in the vector sum direction and the filtered acceleration data in the three axial directions, and particularly, combining the calculated acceleration data in the vector sum direction and the filtered acceleration data in the three axial directions after self-adaptive filtering into final filtered motion data.

Based on the above embodiment, the overall process of the motion data quality enhancement method may be divided into two parts, namely, a gravity calibration bias calibration process and an adaptive filtering process, and fig. 4 is an overall flowchart of the motion data bias calibration process provided by the present invention, and fig. 5 is an overall flowchart of the motion data filtering process provided by the present invention, and as shown in fig. 4 and 5, the method specifically includes:

wherein the offset calibration process comprises: firstly, acquiring motion data, wherein the motion data are acceleration data obtained by data acquisition when three axes in a triaxial acceleration sensor point to the gravity direction respectively; then, determining three axial calibration functions, and determining three axial calibration values based on the calibration functions and the motion data; then, based on the calibration value and the gravity acceleration in the gravity direction, carrying out parameter optimization on the calibration function to obtain an optimal calibration function; and then, based on the optimal calibration functions in the three axial directions, carrying out offset calibration on acceleration data in the three axial directions in the motion data to obtain calibration motion data.

The calibration function comprises offset and calibration sensitivity, and based on the calibration function and the motion data, calibration values in three axial directions are determined; based on the calibration value and the gravitational acceleration in the gravitational direction, performing parameter optimization on the calibration function to obtain an optimal calibration function, including: determining calibration values in three axial directions based on the motion data and initial values of the calibration function and its offset and calibration sensitivity; determining an optimal value of bias and calibration sensitivity in a calibration function in three axial directions based on the calibration value and the gravitational acceleration; and carrying out parameter optimization on the offset and the calibration sensitivity in the calibration functions in the three axial directions based on the optimal values to obtain the optimal calibration function.

Further, based on the above embodiment, determining the optimum values of the offset and the calibration sensitivity in the calibration function in the three axial directions based on the calibration values, and the gravitational acceleration, includes: residual calculation is carried out based on the calibration value and the gravity acceleration, so as to obtain residual values in three axial directions; and (3) carrying out parameter optimization by a least square method based on the residual value to obtain the optimal values of offset and calibration sensitivity in the calibration functions in three axial directions.

The adaptive filtering process includes: firstly, determining a first array corresponding to acceleration data in three axial directions based on acceleration data in the three axial directions in the calibration motion data; then, equal-length segmentation is carried out on the acceleration data in the first array to obtain a plurality of subarrays; then, based on the range of each subarray, the acceleration data in the first array is subjected to filtering processing to obtain filtered acceleration data; the filtered motion data is then determined based on the filtered acceleration data in the three axes and the acceleration data in the three axes in the calibration motion data.

The method for filtering the acceleration data in the first array based on the range of each subarray to obtain filtered acceleration data comprises the following steps: under the condition that the range of any subarray is smaller than or equal to a difference threshold value, determining target acceleration data corresponding to the acceleration data in the subarray from acceleration data in a second array, wherein the acceleration data in the second array is obtained by filtering based on the acceleration data in the first array; replacing the acceleration data in the subarray with the target acceleration data to obtain the filtered acceleration data corresponding to the subarray; under the condition that the range of any subarray is larger than a difference value threshold value, determining the acceleration data in the subarray as the filtered acceleration data corresponding to the subarray; and determining the corresponding axial filtered acceleration data based on the filtered acceleration data corresponding to each subarray.

Further, determining filtered motion data based on the filtered acceleration data in the three axes and the acceleration data in the three axes of the calibration motion data, includes: determining acceleration data in a vector sum direction based on acceleration data in three axes in the calibration motion data; the filtered motion data is determined based on the acceleration data in the vector sum direction and the filtered acceleration data in the three axes.

According to the method provided by the embodiment of the invention, the offset calibration is carried out on the motion data, and the self-adaptive filtering processing is carried out on the calibrated motion data, so that the defects of large error, noise inclusion and low quality of the acquired motion data caused by hardware difference in the traditional scheme are overcome, the sampling error is reduced, the noise component is removed, the data quality of the motion data is improved, the quality problem of the motion data can be effectively improved on the premise of limited hardware level, and a foundation is laid for smooth development of motion monitoring and improvement of the motion monitoring accuracy.

The quality enhancement device for motion data provided by the invention is described below, and the quality enhancement device for motion data described below and the quality enhancement method for motion data described above can be referred to correspondingly.

Fig. 6 is a schematic structural diagram of a quality enhancement device for motion data according to the present invention, as shown in fig. 6, the device includes:

an acquiring unit 610, configured to acquire motion data, where the motion data is acceleration data obtained by performing data acquisition when three axes in the triaxial acceleration sensor are respectively directed in a gravity direction;

a calculation unit 620 configured to determine calibration functions in the three axial directions, and determine calibration values in the three axial directions based on the calibration functions and the motion data;

an optimizing unit 630, configured to perform parameter optimization on the calibration function based on the calibration value and the gravitational acceleration in the gravitational direction, so as to obtain an optimal calibration function;

and the calibration unit 640 is configured to perform offset calibration on the acceleration data in the three axial directions in the motion data based on the optimal calibration functions in the three axial directions, so as to obtain calibration motion data.

According to the quality enhancement device for the motion data, the three axial calibration values are determined through the three axial calibration functions and the motion data, and the motion data are acceleration data obtained by data acquisition when the three axes in the triaxial acceleration sensor point to the gravity direction respectively; according to the calibration value and the gravity acceleration in the gravity direction, parameter optimization is carried out on the calibration function, bias calibration is carried out on acceleration data in three axial directions in the motion data according to the optimized optimal calibration function, so that the calibration motion data are obtained, the quality improvement of the motion data is realized, the defects of large error and low quality of the acquired motion data caused by hardware difference in the traditional scheme are overcome, the sampling error is reduced through the bias calibration of the motion data, the data quality is improved, the quality problem of the motion data can be effectively improved on the premise of limited hardware level, and a foundation is laid for smooth development of motion monitoring and improvement of the motion monitoring accuracy.

Based on the above embodiment, the calibration function includes bias and calibration sensitivity;

the optimizing unit 630 is configured to:

Based on the above embodiment, the optimizing unit 630 is configured to:

Based on the above embodiment, the apparatus further comprises a filtering unit configured to:

Based on the above embodiments, the filtering unit is configured to:

FIG. 7 is a schematic diagram of a motion data quality enhancement system according to the present invention, as shown in FIG. 7, which includes a tri-axial acceleration sensor 710, and a motion data quality enhancement device 720;

the triaxial acceleration sensor 710 is configured to acquire data when three axes are respectively pointing to a gravity direction, obtain motion data, and transmit the motion data to the quality enhancement device 720 of the motion data;

the quality enhancement device 720 of the motion data is configured to obtain the motion data, determine calibration values of the three axes based on the motion data and calibration functions of the three axes, optimize parameters of the calibration functions based on the calibration values and gravitational acceleration of the gravitational direction to obtain an optimal calibration function, and offset calibrate acceleration data of the three axes in the motion data based on the optimal calibration function of the three axes to obtain calibration motion data.

The invention provides a quality enhancement system of motion data, which comprises a triaxial acceleration sensor and a quality enhancement device of the motion data, wherein the triaxial acceleration sensor is used for acquiring the motion data and transmitting the motion data to the quality enhancement device of the motion data; the quality enhancement device of the motion data determines three axial calibration values based on the motion data and three axial calibration functions, performs parameter optimization on the calibration functions based on the calibration values and the gravity acceleration in the gravity direction to obtain an optimal calibration function, performs offset calibration on the acceleration data in the three axial directions in the motion data based on the three axial optimal calibration functions to obtain calibration motion data, achieves quality improvement of the motion data, overcomes the defects of large error and low quality of the motion data acquired due to hardware difference, reduces sampling error through offset calibration of the motion data, improves the quality of the motion data, and effectively improves the quality problem of the motion data on the premise of limited hardware level, and lays a foundation for smooth development of motion monitoring and improvement of motion monitoring accuracy.

Fig. 8 illustrates a physical structure diagram of an electronic device, as shown in fig. 8, which may include: processor 810, communication interface 820, memory 830, and communication bus 840, wherein processor 810, communication interface 820, and memory 830 accomplish communication with each other through communication bus 840. Processor 810 may invoke logic instructions in memory 830 to perform a quality enhancement method for motion data, the method comprising: acquiring motion data, wherein the motion data are acceleration data obtained by data acquisition when three axes in a triaxial acceleration sensor point to the gravity direction respectively; determining calibration functions in the three axial directions, and determining calibration values in the three axial directions based on the calibration functions and the motion data; based on the calibration value and the gravity acceleration in the gravity direction, performing parameter optimization on the calibration function to obtain an optimal calibration function; and carrying out offset calibration on the acceleration data in the three axial directions in the motion data based on the optimal calibration functions in the three axial directions to obtain calibration motion data.

Further, the logic instructions in the memory 830 described above may be implemented in the form of software functional units and may be stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-only memory (ROM), a random access memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

In another aspect, the present invention also provides a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, enable the computer to perform a method of enhancing the quality of motion data provided by the above methods, the method comprising: acquiring motion data, wherein the motion data are acceleration data obtained by data acquisition when three axes in a triaxial acceleration sensor point to the gravity direction respectively; determining calibration functions in the three axial directions, and determining calibration values in the three axial directions based on the calibration functions and the motion data; based on the calibration value and the gravity acceleration in the gravity direction, performing parameter optimization on the calibration function to obtain an optimal calibration function; and carrying out offset calibration on the acceleration data in the three axial directions in the motion data based on the optimal calibration functions in the three axial directions to obtain calibration motion data.

In yet another aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, is implemented to perform a method of enhancing quality of motion data provided by the methods described above, the method comprising: acquiring motion data, wherein the motion data are acceleration data obtained by data acquisition when three axes in a triaxial acceleration sensor point to the gravity direction respectively; determining calibration functions in the three axial directions, and determining calibration values in the three axial directions based on the calibration functions and the motion data; based on the calibration value and the gravity acceleration in the gravity direction, performing parameter optimization on the calibration function to obtain an optimal calibration function; and carrying out offset calibration on the acceleration data in the three axial directions in the motion data based on the optimal calibration functions in the three axial directions to obtain calibration motion data.

The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.

From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method of quality enhancement of motion data, comprising:

2. The method of claim 1, wherein the calibration function includes bias and calibration sensitivity;

3. The method of claim 2, wherein determining an optimal value of offset and calibration sensitivity in the three-axis calibration function based on the calibration values and the gravitational acceleration comprises:

4. A method of enhancing the quality of motion data according to any one of claims 1 to 3, wherein said performing offset calibration on acceleration data in said three axial directions in said motion data based on said three axial directions optimal calibration function, to obtain calibrated motion data, further comprises:

5. The method of claim 4, wherein filtering the acceleration data in the first array based on the range of each sub-array to obtain filtered acceleration data, comprising:

under the condition that the range of any one of the subarrays is smaller than or equal to a difference threshold value, determining target acceleration data corresponding to the acceleration data in any one of the subarrays from acceleration data in a second array, wherein the acceleration data in the second array is obtained by filtering processing based on the acceleration data in the first array;

6. The method of claim 4, wherein determining filtered motion data based on the filtered acceleration data in the three axes and the acceleration data in the three axes of the calibration motion data comprises:

7. A quality enhancement device for motion data, comprising:

8. The quality enhancement system of the motion data is characterized by comprising a triaxial acceleration sensor and a quality enhancement device of the motion data;

9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements a method of quality enhancement of motion data as claimed in any one of claims 1 to 6 when the program is executed by the processor.

10. A non-transitory computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when executed by a processor, implements a quality enhancement method of motion data according to any one of claims 1 to 6.