An adaptive data collection method and system for smart wearable devices

By analyzing the status monitoring sequence of the smart wearable device, distinguishing and processing extreme values, and determining the optimal interpolation point, the problem that inherent modal functions are not conducive to compression is solved, and data compression efficiency and transmission efficiency are improved.

CN117728842BActive Publication Date: 2024-11-29深圳市阿龙电子有限公司
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410022395.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-01-08
Publication Date
2024-11-29
Estimated Expiration
2044-01-08

AI Technical Summary

Technical Problem

The behavioral measurement data collected by smart wearable devices is manifested as long-term non-stationary periodic fluctuations, which leads to the acquired inherent modal functions that are not conducive to compression of the compression algorithm and affects the compression efficiency of the monitoring data.

Method used

By obtaining the status monitoring sequence of smart wearable devices, analyzing the degree of non-labelable extreme value points of similar local extreme value points, distinguishing between labelable and non-labelable extreme value points, determining the optimal interpolation point based on the position and distribution importance of unlabelable characteristic points, and improving the package The smoothness of the network line, resulting in an inherent modal function that is more suitable for compression.

Benefits of technology

Without losing user behavior characteristic information, data compression efficiency is improved, envelope smoothness is enhanced, and data transmission efficiency is improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117728842B_ABST
    Figure CN117728842B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of data compression technology, and specifically to an adaptive data acquisition method and system for smart wearable devices; obtaining markable extreme points and unmarkable extreme points according to data difference characteristics between similar local extreme points in a state monitoring sequence; obtaining an unmarkable data group according to the distribution characteristics of the markable extreme points and the unmarkable extreme points; obtaining the final importance and optimal interpolation points of the unmarkable feature points according to the data interval characteristics and distribution characteristics in the unmarkable data group. The present invention obtains an envelope and an intrinsic mode function according to the markable extreme points and the optimal interpolation points; calculates the energy value of the intrinsic mode function, compresses and transmits the intrinsic mode function whose energy value exceeds a preset energy threshold through a compression algorithm, and improves data compression efficiency without losing user behavior feature information.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of data compression, and in particular to an adaptive data acquisition method and system for an intelligent wearable device. Background Art

[0002] Smart wearable devices are portable devices that integrate multiple sensors and computing capabilities. They are usually connected to users' mobile phones, computers and other devices through wireless technology to transmit data and receive instructions in real time. Since smart wearable devices collect a large number of data types and collect them frequently, in order to ensure the efficiency of data transmission, the collected data needs to be compressed.

[0003] Combined with the characteristics of user behavior measurement data in wearable device scenarios, the existing EMD empirical mode decomposition algorithm can be used to compress the monitoring data. The principle is to decompose the pre-processed monitoring data into different intrinsic mode functions through the algorithm and retain the intrinsic mode functions with larger energy, and then compress them through the existing compression algorithm to improve the compression efficiency. However, since the behavior measurement data collected by wearable devices exhibits long-term non-stationary periodic fluctuation characteristics and random changes within a certain range, the envelope obtained by the cubic spline interpolation method is more complex, resulting in unnecessary fluctuations, resulting in the acquired intrinsic mode function being not conducive to compression by the compression algorithm, thereby affecting the compression efficiency of the monitoring data of smart wearable devices. Summary of the invention

[0004] In order to solve the technical problem that the intrinsic mode function obtained above is not conducive to compression by the compression algorithm, affecting the compression efficiency of the monitoring data of the smart wearable device, the purpose of the present invention is to provide an adaptive data acquisition method and system for a smart wearable device, and the technical solution adopted is as follows:

[0005] Obtain the status monitoring sequence of smart wearable devices;

[0006] Obtaining the unmarkable degree of local extreme value points of the same type according to the data difference characteristics between the local extreme value points of the same type in the state monitoring sequence; obtaining markable extreme value points and unmarkable extreme value points according to the unmarkable degree; obtaining unmarkable feature points and unmarkable data groups according to the distribution characteristics of markable extreme value points and unmarkable extreme points;

[0007] Obtaining the position importance of the unmarkable feature points according to the data interval characteristics in the unmarkable data group; obtaining the distribution importance of the unmarkable feature points according to the data distribution characteristics in the unmarkable data group; obtaining the final importance of the unmarkable feature points according to the position importance and the distribution importance; obtaining the optimal interpolation point according to the final importance of the unmarkable feature points in the unmarkable data group;

[0008] An envelope is obtained by cubic spline interpolation based on the markable extreme points and the optimal interpolation points; an intrinsic mode function is obtained by empirical mode decomposition based on the envelope; the energy value of the intrinsic mode function is calculated, and the intrinsic mode function whose energy value exceeds a preset energy threshold is compressed and transmitted through a compression algorithm.

[0009] Furthermore, the step of obtaining the unmarkability degree of similar local extreme value points according to the data difference characteristics between the similar local extreme value points in the state monitoring sequence includes:

[0010] Mapping the values ​​of any similar local extreme value points in a negative correlation manner to obtain an extreme value representation value; calculating the product of the extreme value representation value and a preset first constant to obtain a first degree;

[0011] Calculate the average value of the absolute value of the difference between any local extreme point of the same type and an adjacent local extreme point of the same type to obtain a local difference average value; calculate the average value of the local difference average values ​​of local extreme points of the same type to obtain an overall difference average value; calculate the ratio of the local difference average value to the overall difference average value and perform negative correlation mapping to obtain a local discrete representation value; calculate the product of the local discrete representation value and a preset second constant to obtain a second degree;

[0012] The sum of the first degree and the second degree of any local extreme value point of the same type is calculated to obtain the unmarkable degree of any local extreme value point of the same type.

[0013] Furthermore, the step of obtaining the markable extreme value points and the unmarkable extreme value points according to the unmarkable degree includes:

[0014] When the unmarkability of the local extreme value points of the same type exceeds a preset first threshold, the local extreme value points of the same type are unmarkable extreme value points, otherwise they are markable extreme value points.

[0015] Furthermore, the step of obtaining unmarkable feature points and unmarkable data groups according to the distribution characteristics of the markable extreme points and the unmarkable extreme points includes:

[0016] For any interval range of adjacent markable extreme points in the state monitoring sequence, when the number of unmarkable extreme points in the interval range exceeds a preset second threshold, the unmarkable extreme points in the interval range are split according to a preset splitting condition to obtain different unmarkable data groups, and the unmarkable extreme points in the unmarkable data groups are unmarkable feature points; the preset splitting condition is that the number of unmarkable extreme points of each of the unmarkable data groups is within the range of a preset number interval, and the number of unmarkable data groups obtained by splitting is the largest.

[0017] Furthermore, the step of obtaining the position importance of the unmarkable feature point according to the data interval feature in the unmarkable data group includes:

[0018] Construct a two-dimensional rectangular coordinate system of the state monitoring sequence, wherein the horizontal axis of the two-dimensional rectangular coordinate system is the timestamp of the data point in the state monitoring sequence, and the vertical axis is the numerical value of the data point; calculate the average value of the timestamps of the unmarkable feature points in the unmarkable data group to obtain the central timestamp; calculate the absolute value of the difference between the timestamp of the unmarkable feature point in the unmarkable data group and the central timestamp and normalize it to obtain the position importance of the unmarkable feature point.

[0019] Furthermore, the step of obtaining the distribution importance of the unmarkable feature points according to the data distribution characteristics in the unmarkable data group includes:

[0020] In the two-dimensional rectangular coordinate system, the maximum value of the Euclidean distance between the unmarkable feature point and the adjacent unmarkable feature point is calculated and normalized to obtain an adjacent distance feature value; the product of the adjacent distance feature value and a preset third constant is calculated to obtain a first distribution value; the numerical value of the unmarkable feature point is normalized to obtain a data characterization value; the product of the data characterization value and a preset fourth constant is calculated to obtain a second distribution value; the sum of the first distribution value and the second distribution value is calculated to obtain the distribution importance of the unmarkable feature point.

[0021] Furthermore, the step of obtaining the final importance of the unmarkable feature point according to the position importance and the distribution importance includes:

[0022] An average value of the position importance and the distribution importance is calculated to obtain a final importance of the unmarkable feature point.

[0023] Furthermore, the step of obtaining the optimal interpolation point according to the final importance of the unmarkable feature points in the unmarkable data group includes:

[0024] According to the final importance of the unmarkable feature points in the unmarkable data group, weights are assigned to the unmarkable feature points in the unmarkable data group, and the greater the final importance of the unmarkable feature point, the greater the corresponding weight; the average value of the product of the unmarkable feature points in the unmarkable data group and the corresponding weight is calculated to obtain the optimal interpolation value; the optimal interpolation point of the unmarkable data group is obtained according to the optimal interpolation value and the central timestamp.

[0025] The present invention also proposes an adaptive data acquisition system for a smart wearable device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any one of the steps of an adaptive data acquisition method for a smart wearable device.

[0026] The present invention has the following beneficial effects:

[0027] In the present invention, obtaining the unmarkable degree of similar local extreme points can reflect the degree of representation of user behavior information. Obtaining markable extreme points and unmarkable extreme points according to the unmarkable degree can determine the extreme points involved in obtaining the envelope line; obtaining unmarkable feature points and unmarkable data groups can further analyze the unmarkable data points, and obtain the optimal interpolation points while retaining the data features, thereby increasing the smoothness of the envelope line. Obtaining the position importance can give a greater weight to the unmarkable feature points that can better represent the user behavior information according to the data interval characteristics of the unmarkable feature points; obtaining the distribution importance can give a greater weight to the unmarkable feature points that can better represent the user behavior information according to the data distribution characteristics of the unmarkable feature points; and then obtaining the optimal interpolation points according to the final importance. Finally, the envelope line is obtained according to the markable extreme points and the optimal interpolation points, which can increase the smoothness of the envelope line; so that the data compression efficiency is improved without losing the user behavior feature information. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0029] Figure 1 A flow chart of an adaptive data collection method for a smart wearable device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the adaptive data acquisition method and system of a smart wearable device proposed by the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0031] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0032] The following is a detailed description of a method and system for adaptive data collection for a smart wearable device provided by the present invention in conjunction with the accompanying drawings.

[0033] See also Figure 1 , which shows a flow chart of an adaptive data collection method for a smart wearable device provided by an embodiment of the present invention, the method comprising the following steps:

[0034] Step S1, obtaining a status monitoring sequence of a smart wearable device.

[0035] In an embodiment of the present invention, the implementation scenario is to compress and transmit the monitoring data of the smart wearable device; first, the state monitoring sequence of the smart wearable device is obtained. The embodiment of the present invention takes heart rate data as an example, and the processing steps of other collection objects are the same. The heart rate data time series collected by the smart device is denoised and preprocessed to obtain the state monitoring sequence of the heart rate data. It should be noted that the implementer can determine the collection object according to the implementation scenario. Denoising belongs to the existing technology, and the specific denoising steps will not be repeated.

[0036] Step S2, obtaining the unmarkability of similar local extreme points according to the data difference characteristics between similar local extreme points in the state monitoring sequence; obtaining markable extreme points and unmarkable extreme points according to the unmarkability; obtaining unmarkable feature points and unmarkable data groups according to the distribution characteristics of markable extreme points and unmarkable extreme points.

[0037] After obtaining the state monitoring sequence, the system will generate a large amount of data due to the high acquisition frequency, so it is necessary to compress and transmit the state monitoring sequence. During the compression process, the state monitoring sequence is subjected to the EMD empirical mode decomposition algorithm to obtain multiple intrinsic mode functions, and the intrinsic mode functions with larger energy are selected and compressed through the compression algorithm, thereby improving the storage and transmission efficiency of the state monitoring sequence. It should be noted that the EMD algorithm belongs to the prior art, and the specific decomposition steps are not repeated. In the EMD algorithm, the acquisition of the upper and lower envelopes has an important influence on the decomposition process of the intrinsic mode function and the final compression processing result; since the behavioral measurement data collected by the wearable device exhibits long-term non-stationary periodic fluctuation characteristics, random changes may occur within a certain range, and local noise is not removed, resulting in the acquisition of the intrinsic mode function. The envelope is not smooth and the complexity is high, which reduces the compression efficiency of the intrinsic mode function and is not conducive to data transmission; therefore, it is necessary to improve the acquisition of the envelope so that the envelope and the intrinsic mode function are more suitable for compression by the compression algorithm.

[0038] The generation of the envelope is achieved by marking the local maximum or minimum points in the state monitoring sequence and connecting them using the cubic spline interpolation method. However, for the data sequence of user behavior characteristics obtained in the implementation scenario, there are a large number of dense local oscillation extreme points and some unremoved noise points. Therefore, it is necessary to analyze the characteristics of the extreme points and select the necessary extreme points to construct the envelope, reduce unnecessary fluctuations of the envelope, and improve the compression efficiency of the intrinsic mode function. First, the unmarkable degree of similar local extreme points can be obtained based on the data difference characteristics between similar local extreme points in the state monitoring sequence.

[0039] Preferably, in one embodiment of the present invention, the step of obtaining the unmarkable degree includes: mapping the numerical values ​​of any similar local extreme value points in a negative correlation to obtain an extreme value characterization value; calculating the product of the extreme value characterization value and a preset first constant to obtain a first degree; similar local extreme value points refer to all local maximum value points or all local minimum value points. For example, when the lower envelope line is obtained, only all local minimum value points are analyzed, and similar local extreme value points are local minimum value points; in the embodiment of the present invention, taking the upper envelope line as an example, similar local extreme value points are local maximum value points; the acquisition process of the lower envelope line is the same. When the value of any similar local extreme value point is smaller, the extreme value characterization value is larger, which means that the extreme value characteristics of any similar local extreme value point in the local maximum value point in the state monitoring sequence are less obvious, and the first degree is larger. In obtaining the envelope, the arbitrary similar local extreme value point can be discarded to increase smoothness; when the value of any similar local extreme value point is larger, the extreme value characterization value is smaller, which means that the extreme value characteristics of any similar local extreme value point in the local maximum value point in the state monitoring sequence are more obvious, and the first degree is smaller. The arbitrary similar local extreme value point can more reflect the changes in the user's behavioral characteristics, and in obtaining the envelope, the arbitrary similar local extreme value point cannot be discarded. The purpose of presetting the first constant is to serve as a weight. In the embodiment of the present invention, it is 0.4, and the implementer can determine it according to the implementation scenario. It should be noted that in obtaining the lower envelope, the extreme value representation value is obtained by normalizing the numerical values ​​of similar local extreme points, because the larger the numerical value of the local minimum point, the less obvious the extreme value characteristics of the similar local extreme point in the local minimum point, and the more it can be discarded in obtaining the lower envelope.

[0040] Further, the average value of the absolute value of the difference between any local extreme point of the same type and two adjacent local extreme points of the same type on the left and right is calculated to obtain the local difference average value; the local extreme points of the same type at both ends only calculate the absolute value of the difference with one adjacent local extreme point of the same type as the local difference average value; the average value of the local difference average values ​​of the local extreme points of the same type is calculated to obtain the overall difference average value. The ratio of the local difference average value to the overall difference average value is calculated and negatively correlated to obtain a local discrete representation value; the product of the local discrete representation value and the preset second constant is calculated to obtain the second degree. When the local difference average value is smaller than the overall difference average value, it means that the fluctuation characteristics at any local extreme point of the same type are less obvious, the local discrete representation value is larger, the second degree is larger, and the any local extreme point of the same type can be discarded in obtaining the envelope; on the contrary, when the local difference average value is larger than the overall difference average value, the local discrete representation value is smaller, the second degree is smaller, which means that the fluctuation characteristics at any local extreme point of the same type are more obvious, the change in the user's behavioral characteristics is more obvious, and in obtaining the envelope, the any local extreme point of the same type cannot be discarded. The purpose of presetting the second constant is to serve as a weight. In the embodiment of the present invention, it is 0.6, and the implementer can determine it according to the implementation scenario. Calculate the sum of the first degree and the second degree of any local extreme value point of the same type to obtain the unmarkable degree of any local extreme value point of the same type; the greater the unmarkable degree, the more likely the local extreme value point of the same type can be discarded in the process of obtaining the envelope; the formula for obtaining the unmarkable degree includes:

[0041]

[0042] In the formula, Indicates the degree of unmarkability of any similar local extreme point, Indicates the preset first constant, Indicates the preset second constant, represents the value of any local extreme point of the same type, represents an exponential function with a natural constant as base, represents the local difference average value of any similar local extreme point, represents the number of similar local extreme points, Indicates The average value of the local differences of similar local extreme points, represents the overall difference mean; represents the extreme value, Indicates the first degree, represents the local discrete representation value, Indicates the second degree.

[0043] Further, the markable extreme points and unmarkable extreme points are obtained according to the unmarkable degree, specifically including: when the unmarkable degree of the same type of local extreme point exceeds the preset first threshold, the same type of local extreme point is an unmarkable extreme point, otherwise it is a markable extreme point; in the embodiment of the present invention, the preset first threshold is 0.5, and the implementer can determine it according to the implementation scenario. The markable extreme point can directly participate in the acquisition of the envelope, while for the unmarkable extreme point, further analysis is required, and for multiple adjacent unmarkable extreme points, it is necessary to select appropriate interpolation points from them to participate in the acquisition of the envelope.

[0044] For unmarkable extreme points, grouping processing is required. According to the characteristics of the unmarkable extreme points in the group, analysis is performed to obtain suitable interpolation points, thereby reducing the number of interpolation points and eliminating the influence of noise. While retaining the changing trend of the user's behavioral characteristics in the original data, the compression efficiency is improved. Therefore, the unmarkable feature points and unmarkable data groups are obtained according to the distribution characteristics of the markable extreme points and the unmarkable extreme points.

[0045] Preferably, in one embodiment of the present invention, the step of obtaining unmarkable feature points and unmarkable data groups includes: for any interval range of adjacent markable extreme points in the state monitoring sequence, that is, taking any two adjacent markable extreme points as boundary points to obtain the interval range, when the number of unmarkable extreme points in the interval range exceeds a preset second threshold, splitting the unmarkable extreme points in the interval range according to a preset splitting condition to obtain different unmarkable data groups, and the unmarkable extreme points in the unmarkable data groups are unmarkable feature points; the preset splitting condition is that the number of unmarkable extreme points of each unmarkable data group is within the range of a preset number interval, and the number of unmarkable data groups obtained by splitting is the largest. In the embodiment of the present invention, the second threshold is preset to 2. If the number of unmarkable extreme value points within the interval does not exceed 2, the unmarkable extreme value points within the interval are discarded. Since the number of unmarkable extreme value points is too small, the influence on the acquisition of the envelope line can be ignored. The preset data interval is 3 to 5, the number of unmarkable extreme value points in each unmarkable data group is not less than 3 and not more than 5, and it is necessary to ensure that the number of unmarkable data groups obtained by splitting is the largest. For example, if there are 13 unmarkable extreme value points in the interval, the number of unmarkable feature points in each unmarkable data group obtained by splitting according to the order of the state monitoring sequence is respectively Obtaining similar numbers of unlabeled data groups enables selecting appropriate interpolation points in each unlabeled data group and making the distances between the interpolation points close, thus maintaining the smoothness and accuracy of the envelope.

[0046] Step S3, obtaining the position importance of the unmarkable feature points according to the data interval characteristics in the unmarkable data group; obtaining the distribution importance of the unmarkable feature points according to the data distribution characteristics in the unmarkable data group; obtaining the final importance of the unmarkable feature points according to the position importance and the distribution importance; obtaining the optimal interpolation point according to the final importance of the unmarkable feature points in the unmarkable data group.

[0047] After obtaining the unmarkable data group and the unmarkable feature point, the position importance of the unmarkable feature point can be obtained according to the data interval characteristics in the unmarkable data group. Preferably, in one embodiment of the present invention, obtaining the position importance includes: constructing a two-dimensional rectangular coordinate system of the state monitoring sequence, the horizontal axis of the two-dimensional rectangular coordinate system is the timestamp of the data point in the state monitoring sequence, and the vertical axis is the value of the data point; calculating the average value of the timestamps of the unmarkable feature points in the unmarkable data group to obtain the center timestamp; calculating the absolute value of the difference between the timestamp of the unmarkable feature point in the unmarkable data group and the center timestamp and normalizing it to obtain the position importance of the unmarkable feature point; according to the heart rate data collected in the implementation scenario, it can be known that if the unmarkable feature point appears in the early or late stage of a range, it is more likely to contain more start or end information about the user behavior characteristics than other extreme points. Therefore, when the timestamp of the unmarkable feature point is farther away from the center timestamp, the more user behavior information represented by the unmarkable feature point, the more important the unmarkable feature point is, and the greater the position importance.

[0048] Furthermore, the data difference characteristics between the unmarkable feature points can also characterize the changing trend of user behavior, so the distribution importance of the unmarkable feature points can be obtained according to the data distribution characteristics in the unmarkable data group; preferably, in one embodiment of the present invention, the step of obtaining the distribution importance includes: in the two-dimensional rectangular coordinate system, the maximum value of the Euclidean distance between the unmarkable feature point and the two adjacent unmarkable feature points on the left and right is calculated and normalized to obtain the adjacent distance feature value; the larger the adjacent distance feature value, the farther the distance between the two unmarkable feature points. Calculate the product of the adjacent distance feature value and the preset third constant to obtain the first distribution value; in the embodiment of the present invention, the preset third constant is 0.6, the purpose is to assign weights, and the implementer can determine it according to the implementation scenario; when the adjacent distance feature value of the unmarkable feature point is larger, it means that the distance from the adjacent unmarkable feature point is farther, the position of the unmarkable feature point is more special, and it can reflect the data characteristics under the abnormal state of user behavior, and the unmarkable feature point is more important; on the contrary, if the adjacent distance feature value is smaller, it means that the distribution is denser and the importance is lower. Normalize the values ​​of the unmarkable feature points to obtain the data characterization value; the larger the data characterization value, the more prominent the value is, and the more it can reflect the user's behavior information; it should be noted that in the process of calculating the lower envelope, the data characterization value is obtained by mapping the values ​​of the unmarkable feature points in a negative correlation, because the smaller the data of the local minimum point, the more it can reflect the user's behavior information. Calculate the product of the data characterization value and the preset fourth constant to obtain the second distribution value; in an embodiment of the present invention, the preset fourth constant is 0.4, and the implementer can determine it according to the implementation scenario. Calculate the sum of the first distribution value and the second distribution value to obtain the distribution importance of the unmarkable feature point; the formula for obtaining the distribution importance includes:

[0049]

[0050] In the formula, Represents the distribution importance of unmarkable feature points, Indicates the preset third constant, Indicates the preset fourth constant, Indicates the value of the unmarkable feature point. represents an exponential function with a natural constant as base, Indicates the adjacent distance feature value of the unmarkable feature point, represents the normalization function, represents the first distribution value, Represents the second distribution value.

[0051] Furthermore, the final importance of the unmarkable feature point can be obtained according to the position importance and the distribution importance, specifically including: calculating the average of the position importance and the distribution importance to obtain the final importance of the unmarkable feature point. The greater the final importance of the unmarkable feature point, the greater the importance of the user behavior information. Then, the optimal interpolation point can be obtained according to the final importance of the unmarkable feature point in the unmarkable data group.

[0052] Preferably, in one embodiment of the present invention, the step of obtaining the optimal interpolation point includes: assigning weights to the unmarkable feature points in the unmarkable data group according to the final importance of the unmarkable feature points in the unmarkable data group, wherein the greater the final importance of the unmarkable feature point, the greater the corresponding weight. In the embodiment of the present invention, the sum of the weights of the unmarkable feature points in the unmarkable data group is 1. For example, the weights assigned to the unmarkable feature points in the unmarkable data group according to the size of the final importance are respectively: , implementers can determine it based on the implementation scenario. Calculate the average value of the product of the unmarkable feature points and the corresponding weights in the unmarkable data group to obtain the optimal interpolation value; obtain the optimal interpolation point of the unmarkable data group based on the optimal interpolation value and the center timestamp, that is, the horizontal axis coordinate of the optimal interpolation point is the center timestamp of the unmarkable data group, and the vertical axis coordinate is the optimal interpolation value. The optimal interpolation point reflects the overall characteristics of the unmarkable feature points in the unmarkable data group, reduces the number of interpolation points, and increases the smoothness of the envelope.

[0053] Step S4, obtaining an envelope curve through cubic spline interpolation according to the markable extreme points and the optimal interpolation points; obtaining an intrinsic mode function through empirical mode decomposition according to the envelope curve; calculating the energy value of the intrinsic mode function, and compressing and transmitting the intrinsic mode function whose energy value exceeds a preset energy threshold through a compression algorithm.

[0054] Markable extreme points are more important extreme points in the monitoring state sequence, and the optimal interpolation points are points obtained by analyzing multiple unmarkable feature points. Therefore, the envelope can be obtained by cubic spline interpolation based on the markable extreme points and the optimal interpolation points. It should be noted that obtaining the envelope by cubic spline interpolation belongs to the prior art, and the specific calculation steps will not be repeated. Further, the intrinsic mode function can be obtained by empirical mode decomposition based on the envelope. It should be noted that obtaining the intrinsic mode function belongs to the prior art, and the specific acquisition steps will not be repeated. Finally, the energy value of the intrinsic mode function can be calculated, and the intrinsic mode function whose energy value exceeds the preset energy threshold is compressed and transmitted through a compression algorithm; the intrinsic mode function with a larger energy value can better reflect the user behavior information. It should be noted that the calculation step of the energy value of the intrinsic mode function belongs to the prior art, and the specific calculation steps will not be repeated. The preset energy threshold can be determined by the implementer according to the implementation scenario. In the embodiment of the present invention, compression is performed through the existing run-length coding. At this point, by analyzing similar local extreme points in the state monitoring sequence to obtain the optimal interpolation point, the smoothness of the envelope is improved while ensuring that the user behavior feature information is not missing, thereby increasing the compression efficiency.

[0055] In summary, the embodiments of the present invention provide an adaptive data collection method for smart wearable devices; markable extreme points and unmarkable extreme points are obtained according to the data difference characteristics between similar local extreme points in the state monitoring sequence; an unmarkable data group is obtained according to the distribution characteristics of the markable extreme points and the unmarkable extreme points; the final importance and optimal interpolation points of the unmarkable feature points are obtained according to the data interval characteristics and distribution characteristics in the unmarkable data group. The present invention obtains envelope curves and intrinsic mode functions based on the markable extreme points and the optimal interpolation points; calculates the energy value of the intrinsic mode function, compresses and transmits the intrinsic mode function whose energy value exceeds the preset energy threshold through a compression algorithm, and improves data compression efficiency without losing user behavior feature information.

[0056] The present invention also proposes an adaptive data acquisition system for a smart wearable device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any one of the steps of an adaptive data acquisition method for a smart wearable device.

[0057] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0058] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. An adaptive data collection method for a smart wearable device, characterized in that: The method comprises the following steps: Obtain the status monitoring sequence of smart wearable devices; Obtaining the unmarkable degree of local extreme value points of the same type according to the data difference characteristics between the local extreme value points of the same type in the state monitoring sequence; obtaining markable extreme value points and unmarkable extreme value points according to the unmarkable degree; obtaining unmarkable feature points and unmarkable data groups according to the distribution characteristics of markable extreme value points and unmarkable extreme points; Obtaining the position importance of the unmarkable feature points according to the data interval characteristics in the unmarkable data group; obtaining the distribution importance of the unmarkable feature points according to the data distribution characteristics in the unmarkable data group; obtaining the final importance of the unmarkable feature points according to the position importance and the distribution importance; obtaining the optimal interpolation point according to the final importance of the unmarkable feature points in the unmarkable data group; Obtaining an envelope curve by cubic spline interpolation according to the markable extreme point and the optimal interpolation point; obtaining an intrinsic mode function by empirical mode decomposition according to the envelope curve; calculating the energy value of the intrinsic mode function, and compressing and transmitting the intrinsic mode function whose energy value exceeds a preset energy threshold through a compression algorithm; The step of obtaining the unmarkability degree of similar local extreme value points according to the data difference characteristics between the similar local extreme value points in the state monitoring sequence comprises: Mapping the values ​​of any similar local extreme value points in a negative correlation manner to obtain an extreme value representation value; calculating the product of the extreme value representation value and a preset first constant to obtain a first degree; Calculate the average value of the absolute value of the difference between any local extreme point of the same type and an adjacent local extreme point of the same type to obtain a local difference average value; calculate the average value of the local difference average values ​​of local extreme points of the same type to obtain an overall difference average value; calculate the ratio of the local difference average value to the overall difference average value and perform negative correlation mapping to obtain a local discrete representation value; calculate the product of the local discrete representation value and a preset second constant to obtain a second degree; The sum of the first degree and the second degree of any local extreme value point of the same type is calculated to obtain the unmarkable degree of any local extreme value point of the same type.

2. The adaptive data collection method for a smart wearable device according to claim 1, characterized in that: The step of obtaining the markable extreme value points and the unmarkable extreme value points according to the unmarkable degree comprises: When the unmarkability of the local extreme value points of the same type exceeds a preset first threshold, the local extreme value points of the same type are unmarkable extreme value points, otherwise they are markable extreme value points.

3. The adaptive data collection method for a smart wearable device according to claim 1, characterized in that: The step of obtaining unmarkable feature points and unmarkable data groups according to the distribution characteristics of the markable extreme value points and the unmarkable extreme value points comprises: For any interval range of adjacent markable extreme points in the state monitoring sequence, when the number of unmarkable extreme points in the interval range exceeds a preset second threshold, the unmarkable extreme points in the interval range are split according to a preset splitting condition to obtain different unmarkable data groups, and the unmarkable extreme points in the unmarkable data groups are unmarkable feature points; the preset splitting condition is that the number of unmarkable extreme points of each of the unmarkable data groups is within the range of a preset number interval, and the number of unmarkable data groups obtained by splitting is the largest.

4. The adaptive data collection method for a smart wearable device according to claim 1, characterized in that: The step of obtaining the position importance of the unmarkable feature points according to the data interval characteristics in the unmarkable data group comprises: Construct a two-dimensional rectangular coordinate system of the state monitoring sequence, wherein the horizontal axis of the two-dimensional rectangular coordinate system is the timestamp of the data point in the state monitoring sequence, and the vertical axis is the numerical value of the data point; calculate the average value of the timestamps of the unmarkable feature points in the unmarkable data group to obtain the central timestamp; calculate the absolute value of the difference between the timestamp of the unmarkable feature point in the unmarkable data group and the central timestamp and normalize it to obtain the position importance of the unmarkable feature point.

5. The adaptive data collection method for a smart wearable device according to claim 4, characterized in that: The step of obtaining the distribution importance of unmarkable feature points according to the data distribution characteristics in the unmarkable data group comprises: In the two-dimensional rectangular coordinate system, the maximum value of the Euclidean distance between the unmarkable feature point and the adjacent unmarkable feature point is calculated and normalized to obtain an adjacent distance feature value; the product of the adjacent distance feature value and a preset third constant is calculated to obtain a first distribution value; the numerical value of the unmarkable feature point is normalized to obtain a data characterization value; the product of the data characterization value and a preset fourth constant is calculated to obtain a second distribution value; the sum of the first distribution value and the second distribution value is calculated to obtain the distribution importance of the unmarkable feature point.

6. The adaptive data collection method for a smart wearable device according to claim 1, characterized in that: The step of obtaining the final importance of the unmarkable feature point according to the position importance and the distribution importance comprises: An average value of the position importance and the distribution importance is calculated to obtain a final importance of the unmarkable feature point.

7. The adaptive data collection method for a smart wearable device according to claim 4, characterized in that: The step of obtaining the optimal interpolation point according to the final importance of the unmarkable feature points in the unmarkable data group comprises: According to the final importance of the unmarkable feature points in the unmarkable data group, weights are assigned to the unmarkable feature points in the unmarkable data group, and the greater the final importance of the unmarkable feature point, the greater the corresponding weight; the average value of the product of the unmarkable feature points in the unmarkable data group and the corresponding weight is calculated to obtain the optimal interpolation value; the optimal interpolation point of the unmarkable data group is obtained according to the optimal interpolation value and the central timestamp.

8. An adaptive data acquisition system for a smart wearable device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

Citation Information

Patent Citations

  • Data storage method based on user information of smart watch

    CN115670418A

  • Data compression processing method based on big data

    CN117150217A