Posture and sight distance dynamic monitoring system and method based on radio frequency tag identification

Abstract

The present application relates to the field of radio frequency tag identification, and particularly relates to a sitting posture and visual range dynamic monitoring system and method based on radio frequency tag identification, comprising: acquiring user information, presetting an initial polling frequency and an initial collection time length, and collecting visual range data through a reader-writer; decomposing the visual range data to analyze sitting posture adjustment features to adjust the initial collection time length; clustering time based on the fluctuation range of the visual range data and the distribution features of data points in the decomposition result, and adjusting the initial polling frequency according to the clustering processing result; making the reader-writer collect visual range data at the current time by using the adjustment period and the polling frequency of the current time; setting a safe visual range threshold in combination with the user information, and monitoring the sitting posture according to the visual range value at the current time in the visual range data. The present application relies on the advantages of radio frequency identification technology, has the characteristics of non-contact, low power consumption and high adaptability, and realizes a good balance between energy consumption and precision.

Description

Technical Field

[0001] This invention relates to the field of radio frequency tag identification technology, and more specifically to a dynamic monitoring system and method for sitting posture and line of sight based on radio frequency tag identification. Background Technology

[0002] In recent years, the national myopia rate has continued to rise, with poor posture and excessive close viewing distance being the main contributing factors. To address this issue, various vision health monitoring technologies have emerged, among which the method of monitoring posture based on radio frequency identification (RFID) technology is gradually gaining popularity.

[0003] However, existing RFID detection systems have certain drawbacks. First, readers typically emit electromagnetic signals at a constant and relatively high operating frequency to ensure real-time performance, but this leads to over-monitoring when users are relatively still (such as when focused on reading), resulting in wasted battery energy. Conversely, during dynamic user behaviors (such as frequently adjusting posture), emitting electromagnetic signals at a lower operating frequency cannot promptly detect sudden changes in distance, causing warning delays and reducing monitoring accuracy. Therefore, without a behavioral perception mechanism, existing methods cannot effectively balance energy consumption and accuracy. Summary of the Invention

[0004] This invention provides a dynamic monitoring system and method for sitting posture and line of sight based on radio frequency tag identification, in order to solve existing problems.

[0005] The sitting posture and line-of-sight distance dynamic monitoring system and method based on radio frequency tag identification of the present invention adopts the following technical solution:

[0006] One embodiment of the present invention provides a method for dynamic monitoring of sitting posture and line of sight based on radio frequency tag identification, the method comprising the following steps:

[0007] Acquire user information, preset the initial polling frequency and initial collection duration, and collect line-of-sight data through a reader / writer;

[0008] The line-of-sight data is decomposed, and the user's posture adjustment characteristics are analyzed using the data decomposition results. The initial acquisition duration is adjusted to determine the adjustment cycle at the current moment.

[0009] Based on the fluctuation range of the line-of-sight data and the distribution characteristics of the data points in the decomposition results, clustering is performed on all moments in the line-of-sight data. The initial polling frequency is adjusted according to the data distribution of the clustering results and the relative position at the current moment to obtain the polling frequency at the current moment.

[0010] The reader uses the current adjustment cycle and polling frequency to collect the current line-of-sight data;

[0011] Set a safe viewing distance threshold based on user information, and monitor sitting posture based on the viewing distance value at the current moment in the viewing distance data.

[0012] Optionally, the specific methods for decomposing the viewing distance data, analyzing the user's posture adjustment characteristics using the data decomposition results, adjusting the initial acquisition duration, and determining the adjustment cycle at the current moment include:

[0013] The line-of-sight data collected by the reader within the time range corresponding to the initial collection duration after startup, and at the initial polling frequency, will be used as the initial line-of-sight data.

[0014] The STL time series decomposition algorithm is used to decompose the previous line-of-sight data to obtain the trend, periodic and residual terms of the previous line-of-sight data. Based on the data distribution of the trend, periodic and residual terms, the historical regularity coefficient of the user in the previous line-of-sight data is calculated.

[0015] By adjusting the initial data collection duration using historical pattern coefficients, the specific values ​​for the next adjustment cycle can be obtained.

[0016] Optionally, the specific method for obtaining the historical regularity coefficient is as follows:

[0017] Obtain the standard deviation of the slope of all data points in the seasonal term of the previous line-of-sight data. Perform a Fast Fourier Transform on the periodic term of the previous line-of-sight data to obtain the spectrum of the periodic term. Obtain the number of frequencies with non-zero amplitude in the spectrum and calculate the cumulative amplitude value of the corresponding frequencies with non-zero amplitude in the spectrum. Obtain the average value of all data points in the residual term of the previous line-of-sight data. Based on the standard deviation of the slope, the number of frequencies with non-zero amplitude in the spectrum, the cumulative amplitude value, and the average value of all data points in the residual term, obtain the historical regularity coefficient.

[0018] Optionally, determining the adjustment period at the current moment further includes: obtaining the adjustment period at the first... After determining the specific value for the adjustment cycle, when in the first... When at least two adjustment cycles have passed before the first adjustment cycle, according to the first adjustment cycle... The average and standard deviation of the specific values ​​corresponding to all adjustment periods prior to the first adjustment period are used to calculate the first adjustment period. The periodic adjustment coefficient for the first adjustment cycle is used to adjust the periodic adjustment coefficient for the second adjustment cycle. The specific values ​​of the first adjustment cycle are multiplied for adjustment to obtain the adjusted value of the second adjustment cycle. The specific value for each adjustment cycle.

[0019] Optionally, based on the fluctuation range of the line-of-sight data and the distribution characteristics of data points in the decomposition results, clustering is performed on all moments in the line-of-sight data. The initial polling frequency is adjusted according to the data distribution of the clustering results and the relative position at the current moment to obtain the polling frequency at the current moment. The specific method includes:

[0020] By utilizing the discrete characteristics and fluctuations exhibited in the trend, periodic, and residual terms of the previous line-of-sight data, we can analyze the coefficient of change in user posture within the initial acquisition duration range corresponding to the previous line-of-sight data.

[0021] The left neighbor duration is preset, and the sitting posture variation coefficient of any adjustment cycle is combined to obtain the left neighbor duration of the adjustment cycle. Based on the change of viewing distance data at any time in the corresponding left neighbor duration of the adjustment cycle, a sitting posture feature value array formed by several feature values ​​is extracted.

[0022] Clustering is performed on all corresponding time points based on the sitting posture feature value array. The user's sitting posture adjustment rate at the current time is calculated based on the relative position of the current time point in the cluster and the distribution of all time points in the cluster.

[0023] Optionally, the specific method for obtaining the sitting posture variation coefficient is as follows:

[0024] The average information entropy of all STL time-series decomposition results of the previous line-of-sight data and the range of data points in the previous line-of-sight data are obtained. Based on the average information entropy and the range of data points in the previous line-of-sight data, the sitting posture variation coefficient is obtained, wherein the average information entropy and the range of data points in the previous line-of-sight data are both positively correlated with the sitting posture variation coefficient.

[0025] Optionally, the specific method for obtaining the sitting posture feature value array is as follows:

[0026] Preset base left neighbor duration Based on the method for obtaining the sitting posture variation coefficient, the user's sitting posture variation coefficient under any adjustment cycle is calculated, and then... As the left neighbor duration under the corresponding adjustment period, data points in the trend term, period term, and residual term of the spectral distance data within the range of the left neighbor duration under the adjustment period to which the time belongs before any time are obtained. Using the distribution of data points in the trend term, period term, and residual term of the spectral distance data within the range of the left neighbor duration, the trend feature value, period feature value, and residual feature value at the time are calculated respectively. The array formed by the trend feature value, period feature value, and residual feature value at the time is used as the sitting posture feature value array at the time.

[0027] Optionally, the specific method for clustering all corresponding time points based on the sitting posture feature value array includes:

[0028] The Euclidean distance between the sitting posture feature value arrays at different times is used as a distance metric, and the DBSCAN clustering algorithm is used to cluster all times to obtain several clusters. Any cluster is denoted as the sitting posture time cluster.

[0029] Optionally, the specific method for obtaining the sitting posture adjustment rate is as follows:

[0030] If the current moment is an outlier in the cluster space, the posture adjustment rate at the current moment is set to 0. If the current moment is not an outlier in the cluster space, the posture time cluster to which the current moment belongs is obtained, referred to as the current posture time cluster. The standard deviation of the corresponding distances between all moments in the current posture time cluster and the cluster center is obtained, and the average distance between all moments in the current posture time cluster and the cluster center is recorded as the average distance parameter of the current posture time cluster. The corresponding distance between the current moment and the cluster center of the current posture time cluster is recorded as the distance parameter of the current moment. Based on the standard deviation of the corresponding distances between all moments in the current posture time cluster and the cluster center, and the difference between the average distance parameter and the distance parameter of the current moment, the user's posture adjustment rate at the current moment is calculated.

[0031] The sitting posture and line-of-sight dynamic monitoring system based on radio frequency tag identification includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of any one of the sitting posture and line-of-sight dynamic monitoring methods based on radio frequency tag identification.

[0032] The beneficial effects of the technical solution of this invention are as follows: By collecting line-of-sight data and performing in-depth analysis, the system can adaptively identify the user's posture adjustment behavior characteristics, thereby dynamically adjusting the data collection time period and signal polling frequency. When the user is in a stable state, the system automatically extends the collection interval, reduces unnecessary signal transmission, and significantly reduces energy consumption; when frequent posture changes are detected, the sampling density is increased to ensure that key data is not missed, thereby improving the monitoring response speed and accuracy. Through cluster analysis of line-of-sight fluctuation characteristics, the system can accurately determine the current behavioral stage and set a safe distance threshold in combination with individualized user information to achieve personalized real-time early warning. The entire process requires no additional sensors or complex equipment. Relying on the advantages of radio frequency identification technology, it has the characteristics of being non-contact, low-power consumption, and highly adaptable; it achieves a good balance between energy consumption and accuracy, overcomes the shortcomings of traditional monitoring methods such as resource waste in static scenarios and lag in dynamic scenarios, enhances user experience and monitoring reliability, and is suitable for various daily scenarios such as learning and office work. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 This is a flowchart illustrating the steps of the dynamic monitoring method for sitting posture and line of sight based on radio frequency tag identification of the present invention.

[0035] Figure 2 A schematic diagram of line-of-sight monitoring provided in one embodiment of the present invention;

[0036] Figure 3 This is a structural block diagram of the sitting posture and line-of-sight dynamic monitoring system based on radio frequency tag identification of the present invention. Detailed Implementation

[0037] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the dynamic monitoring system and method for sitting posture and line of sight based on RFID tag identification proposed in this invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0038] Unless otherwise defined, 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 pertains.

[0039] The following description, in conjunction with the accompanying drawings, details the specific scheme of the dynamic monitoring system and method for sitting posture and line of sight based on radio frequency tag identification provided by the present invention.

[0040] Please see Figure 1 The diagram illustrates a flowchart of a method for dynamic monitoring of sitting posture and line of sight based on radio frequency tag identification, according to an embodiment of the present invention. The method includes the following steps:

[0041] Step S001: Obtain user information, preset the initial polling frequency and initial collection duration, and collect line-of-sight data through a reader / writer.

[0042] It should be noted that when users are studying or working at their desks, they may unconsciously change their posture due to personal habits, causing their eyes to be too close to the objects they are looking at (such as books, mobile phones, tablets, etc.). This can easily lead to health problems such as hunchback and myopia. Therefore, in order to assist users in monitoring and reminding them of their posture, this embodiment of the invention selects radio frequency tag identification technology to dynamically monitor the posture and viewing distance of users who are studying or working at their desks, and then issue a reminder when the user's posture is incorrect.

[0043] In addition, to facilitate understanding of the user's posture and viewing distance, embodiments of the present invention utilize a smart card with an embedded signal receiving chip and integrated RFID tag reader / writer and Bluetooth functionality, as well as glasses with an embedded RFID tag, to monitor the user's posture and viewing distance. Figure 2 The diagram illustrates visual distance monitoring. During monitoring, the RFID reader in the smart card communicates with the mobile phone via Bluetooth. When the reader sends electromagnetic waves to the RFID tag inside the glasses, the RFID tag's antenna receives the waves and converts them into electrical energy. This electrical energy activates the RFID tag's chip, causing it to work. The chip converts the stored data into electrical signals and sends them back to the reader via the antenna. Upon receiving the signal, the reader uses phase difference detection technology to measure the distance between itself and the RFID tag, obtaining the visual distance data at the corresponding moment. This visual distance data is then transmitted to the mobile app.

[0044] Specifically, in order to implement the dynamic monitoring method for sitting posture and line of sight based on RFID identification proposed in this embodiment, it is first necessary to collect the user's line of sight data and obtain user information. The specific process is as follows:

[0045] Because the reader needs to go through activation, command sending, and data receiving processes during operation, and in work and study scenarios, the distance between the smart card and the smart glasses frame usually keeps the reader in the smart card constantly active, thus continuously sending commands and receiving data, and sending the obtained distance data to the mobile app via Bluetooth. This generates excessive redundant data and also leads to excessive power consumption of the smart card. Therefore, in this embodiment of the invention, an initial polling frequency for the reader is set. and initial collection duration The line-of-sight data collected by the reader within the time range corresponding to the initial collection duration after startup, and at the initial polling frequency, is used as the initial line-of-sight data.

[0046] Preset the initial polling frequency based on experience. and initial collection duration The corresponding values ​​are 20Hz and 5 minutes, respectively.

[0047] Polling frequency is used to represent the process by which the reader completes one acquisition of line-of-sight data at a given time within a unit of time, i.e., a series of actions including activation, sending commands, and receiving data; initial acquisition duration is used to ensure that it can be used by the user from the start.

[0048] In addition, the phone has a mobile app installed, which contains user information, including at least: height, body type, and age.

[0049] Thus, the preset initial polling frequency, initial collection duration, user line-of-sight data, and user information are obtained through the above method.

[0050] Step S002: Decompose the viewing distance data, analyze the user's posture adjustment characteristics using the data decomposition results, and determine the adjustment cycle at the current moment.

[0051] It should be noted that, under normal circumstances, when users are studying or working at a desk, they will adjust their sitting posture multiple times to make themselves more comfortable. This behavior is highly correlated with the user's movement habits. Each time the user adjusts their posture, the viewing distance changes. This means that the user behavior pattern information potentially contained in the posture adjustment behavior is already reflected in the viewing distance data. In order to ensure that the viewing distance data acquired later avoids a large amount of redundant information, this embodiment of the invention chooses to analyze the user's posture adjustment behavior based on the previous viewing distance data in order to adjust the polling frequency of the reader in the smart card.

[0052] In the initial viewing distance data, the fluctuations in viewing distance values ​​contain relevant characteristics of the user's posture and viewing distance adjustment. These characteristics can be used to adjust the polling frequency of the reader in the smart card, so that the polling frequency of the reader can adapt to the rhythm of the user's posture adjustment process during work and study, thereby further reducing the energy consumption required by the reader to transmit and receive signals.

[0053] Specifically, firstly, the STL time series decomposition algorithm is used to decompose the previous line-of-sight data to obtain the trend term, periodic term, and residual term of the previous line-of-sight data. Based on the data distribution in the trend term, periodic term, and residual term, the historical regularity coefficient of the user in the previous line-of-sight data is calculated.

[0054] As a preferred embodiment, the specific method for obtaining the historical regularity coefficient is as follows: obtain the standard deviation of the slope of all data points in the seasonal term of the previous line-of-sight data; perform a fast Fourier transform on the periodic term of the previous line-of-sight data to obtain the spectrum of the periodic term; obtain the number of frequencies with non-zero amplitude in the spectrum; calculate the cumulative amplitude value of the frequencies corresponding to all non-zero amplitudes in the spectrum; obtain the average value of all data points in the residual term of the previous line-of-sight data; and obtain the historical regularity coefficient based on the standard deviation of the slope, the number of frequencies with non-zero amplitude in the spectrum, the cumulative amplitude value, and the average value of all data points in the residual term.

[0055] As an optional embodiment, the specific calculation method for the historical regularity coefficient is as follows: ;in, This represents the historical pattern coefficient of the user in the previous line-of-sight data; The periodic term in the spectrum of the previous line-of-sight data has a non-zero amplitude. The amplitude corresponding to each frequency; This represents the standard deviation of the slope of all data points in the seasonal component of the previous spectral distance data. This represents the number of frequencies with non-zero amplitudes in the spectrum of the periodic term of the previous line-of-sight data. This represents the average value of all data points in the residual term of the previous spectral distance data; This represents an exponential function with the natural constant as its base.

[0056] It should be noted that, for the spectrum of the periodic terms of the early line-of-sight data, in this embodiment of the invention, the periodic terms of the early line-of-sight data are subjected to Fourier transform processing to obtain the corresponding spectrum.

[0057] It should be noted that the regularity coefficient is used to describe the degree of regularity in the user's behavior pattern when adjusting posture in the early stages. The larger the regularity coefficient, the more regular the user's behavior pattern when adjusting posture, and the higher the degree of predictability. In other words, the higher the regularity, the more likely it is to predict future behavior through historical data, thereby reducing unnecessary data collection. Regarding the specific calculation method of the historical regularity coefficient, ... This represents the cumulative amplitude corresponding to frequencies with non-zero amplitudes in the spectrum. A higher number of frequencies with non-zero amplitudes in the spectrum of the sight distance data indicates more frequent and complex user posture adjustments, such as alternating postures. Larger amplitude values ​​further complicate this behavior, reducing the accuracy of predictions regarding user sight distance changes. Additionally, since the seasonal term reflects the periodic characteristics of the corresponding sight distance data, and the slope of the data points in the seasonal term reflects the degree of sight distance change due to posture adjustments within this periodicity, higher dispersion in the seasonal term indicates less regular and more irregular user sight distance adjustments, suggesting frequent and random posture shifts. Finally, the residual term of the previous sight distance data represents the average value of all data points. This value reflects the intensity of random fluctuations. The larger the value, the stronger the unpredictable noise, and the more frequent the sudden changes in the user's sitting posture.

[0058] Then, the initial collection duration is adjusted using historical regularity coefficients to obtain the specific value for the next adjustment cycle.

[0059] As an optional embodiment, the specific value of the next adjustment cycle is calculated as follows: ,in Indicates the first The specific value for each adjustment cycle; Indicates the preset adjustment ratio; Indicates the user's position in the first month. Historical regularity coefficients in the sight distance data within each adjustment cycle; Indicates the first The specific value for each adjustment cycle.

[0060] It should be noted that the preset adjustment ratio is 2 based on experience, and can be adjusted according to the actual situation. This embodiment of the invention does not impose any specific limitations.

[0061] It should be noted that the specific value of the adjustment cycle represents the collection duration of the reader when collecting line-of-sight data. In this embodiment of the invention, the line-of-sight data collection method of the reader is adjusted based on the user's posture adjustment habits during study and office work, thereby obtaining multiple adjustment cycles of the reader during the line-of-sight data collection process. This further enables subsequent data collection in each adjustment cycle to be performed in a corresponding polling frequency adjustment mode, thereby achieving dynamic adjustment of the polling frequency of the reader in the smart card and reducing the power consumption of the smart card. In addition, since the line-of-sight data is collected in real time, the current moment in an adjustment cycle is constantly changing and updating over time. Therefore, to avoid the adjustment cycle from changing with the current moment, The constant changes make it difficult to determine how the adjustment period should end. In this embodiment of the invention, when obtaining the adjustment period, the specific value of the adjustment period at the current moment is obtained only after the end of one adjustment period, and used as the value for the next adjustment period. This value is not obtained again in the next adjustment period. For example, if the current moment is still in the first adjustment period (i.e., the time range corresponding to the initial acquisition duration after the reader starts), the adjustment period is not calculated. However, after the end of the first adjustment period, the specific value of the adjustment period at the current moment is calculated as the specific data acquisition duration for line-of-sight acquisition in the second adjustment period. The calculation of the adjustment period is stopped in the second adjustment period until the end of the second adjustment period, and so on.

[0062] Furthermore, in one embodiment of the present invention: when obtaining the first... After determining the specific value for the adjustment cycle, when in the first... When at least two adjustment cycles have passed before the first adjustment cycle, according to the first adjustment cycle... The average and standard deviation of the specific values ​​corresponding to all adjustment periods prior to the first adjustment period are used to calculate the first adjustment period. The periodic adjustment coefficient for the first adjustment cycle is used to adjust the periodic adjustment coefficient for the second adjustment cycle. The specific values ​​of the first adjustment cycle are multiplied for adjustment to obtain the adjusted value of the second adjustment cycle. The specific value for each adjustment cycle.

[0063] As an optional embodiment, the period adjustment coefficient is calculated as follows: ,in, Indicates the first The periodic adjustment coefficient for each adjustment cycle; Indicates the first The average value corresponding to the specific values ​​of all adjustment cycles preceding the current adjustment cycle; Indicates the first The standard deviation of the specific values ​​corresponding to all adjustment periods prior to the current adjustment period; This represents the sigmoid normalization function.

[0064] It should be noted that the periodic adjustment coefficient is used to describe the periodic adjustment coefficient obtained when the periodic adjustment coefficient is 1. The specific value of each adjustment cycle, and the time the user has experienced. Each adjustment cycle corresponds to a time range within which a user's sitting posture changes within different adjustment cycles. The larger the value of the cycle adjustment coefficient, the more stable the user's posture becomes over the specified period. The more stable the changes in sight distance data within the first adjustment cycle, the better the subsequent determination of the first... When determining the specific value for a given adjustment cycle, the more relevant the values ​​from previous adjustment cycles are, the better. In the specific calculation method for the cycle adjustment coefficient... The larger the value, the longer the average cycle, indicating a more stable time interval for the user to adjust their posture; at the same time, The smaller the value, the smaller the fluctuation in the cycle length of posture adjustment, and the higher the consistency of the user's posture adjustment behavior. The larger the value, the more stable the user's posture adjustment cycle is. Since relying solely on historical regularity coefficients may lead to excessive adjustment of cycle length due to sudden changes in short-term user behavior, this embodiment of the invention introduces a cycle adjustment coefficient to prevent sudden changes in cycle due to a single abnormal data (such as getting up temporarily).

[0065] It should be noted that, in this embodiment of the invention, the adjustment period is used to determine a data collection duration after obtaining the initial viewing distance data. Within this duration, the reader operates at a new polling frequency to ensure that the polling frequency of the series of actions performed by the reader when collecting viewing distance data—including activation, sending commands, and receiving data—matches the changes in the user's posture. For example, if the user frequently adjusts their posture, causing frequent changes in viewing distance and multiple instances of excessively short viewing distances, then a higher polling frequency is needed over a longer period to monitor the user's viewing distance changes and avoid missing viewing distance information during the monitoring process.

[0066] Thus, the adjustment period at the current moment is obtained through the above method.

[0067] Step S003: Based on the fluctuation range of the line-of-sight data and the distribution characteristics of the data points in the decomposition results, perform clustering processing on all time points in the line-of-sight data. According to the data distribution of the clustering results and the relative position at the current time, adjust the initial polling frequency to obtain the polling frequency at the current time.

[0068] It should be noted that after collecting and analyzing the initial viewing distance data, the behavioral pattern characteristics of users adjusting their sitting posture in the current scenario were initially understood, and the adjustment cycle for further viewing distance data collection based on these behavioral pattern characteristics was obtained. In order to ensure that the viewing distance data collected within the adjustment cycle can effectively monitor the user's sitting posture and viewing distance, this invention further utilizes the initial viewing distance data for analysis, thereby obtaining the polling frequency of the reader.

[0069] Specifically, in step S301, the discrete characteristics and fluctuations exhibited in the trend, periodic, and residual terms of the previous line-of-sight data are used to analyze the user's posture variation coefficient within the initial acquisition time range corresponding to the previous line-of-sight data.

[0070] As a preferred embodiment, the specific method for obtaining the sitting posture variation coefficient is as follows: obtain the average information entropy of all STL time-series decomposition results of the previous viewing distance data, and the range of data points in the previous viewing distance data; obtain the sitting posture variation coefficient based on the average information entropy and the range of data points in the previous viewing distance data, wherein the average information entropy and the range of data points in the previous viewing distance data are both positively correlated with the sitting posture variation coefficient.

[0071] As an optional embodiment, the specific calculation method for the sitting posture variation coefficient is as follows: ;in, This represents the coefficient of change in the user's sitting posture within the time range corresponding to the initial viewing distance data; This represents the range of data points in the initial sight distance data; This represents the first [number]th [unit] of the previous sight distance data. Information entropy of each STL time series decomposition result (i.e., trend term, period term, residual term); This represents the sigmoid normalization function.

[0072] It should be noted that the posture variation coefficient describes the overall frequency and magnitude of changes in a user's posture within the data collection period corresponding to the initial visual distance data. A higher posture variation coefficient indicates more frequent and significant changes in posture during the data collection period (i.e., within the initial collection time range corresponding to the initial visual distance data). In the specific calculation method of the posture variation coefficient, the range of data points in the initial visual distance data is considered. The STL decomposition term reflects the magnitude of the change in viewing distance. The information entropy of the STL decomposition term is used to reflect the uncertainty of the data. The larger the entropy value, the stronger the randomness. The overall fluctuation complexity is comprehensively measured by the sum of the information entropy of the three decomposition results. Therefore, this embodiment of the invention provides a basis for dynamically adjusting the polling frequency by quantifying the intensity of the user's sitting posture change during the current collection period. When the viewing distance changes drastically, the monitoring frequency is high, and the frequency is reduced when it is stable.

[0073] Step S302: Preset the basic left neighbor duration and combine it with the sitting posture variation coefficient of any adjustment cycle to obtain the left neighbor duration of the adjustment cycle. Based on the change of viewing distance data at any time in the corresponding left neighbor duration of the adjustment cycle, extract a sitting posture feature value array formed by several feature values.

[0074] As a preferred embodiment, the specific method for obtaining the sitting posture feature value array includes: preset the basic left neighbor duration. Based on the method for obtaining the sitting posture variation coefficient, the user's sitting posture variation coefficient under any adjustment cycle is calculated, and then... As the left neighbor duration under the corresponding adjustment period, data points in the trend term, period term, and residual term of the sight distance data within the corresponding left neighbor duration of the adjustment period to which the time belongs are obtained before any time. Using the distribution of data points in the trend term, period term, and residual term of the sight distance data within the left neighbor duration, the trend feature value, period feature value, and residual feature value at the time are calculated respectively. The array formed by the trend feature value, period feature value, and residual feature value at the time is used as the sitting posture feature value array at the time. This represents the coefficient of change in the user's sitting posture within the time range corresponding to the previous viewing distance data.

[0075] It should be noted that, based on experience, the preset basic left neighbor duration is 10 seconds. Furthermore, since the current moment is the latest moment and the specific values ​​change in real time, the adjustment period to which the current moment belongs is the most recently acquired adjustment period, and a longer period of sight distance data acquisition has not yet been conducted. Therefore, the sight distance data within the corresponding left neighbor duration of the adjustment period at the current moment may be too little. In this embodiment of the invention, to address this situation, the sight distance data from the previous adjustment period to which the current moment belongs is also included within the left neighbor duration range corresponding to the current moment, and the left neighbor duration corresponding to the previous adjustment period is used as the left neighbor duration of the current moment. This method is not used if all data points within the left neighbor duration range of the current moment belong to the adjustment period to which the current moment belongs.

[0076] As an optional embodiment, the specific method for obtaining the trend feature value, periodic feature value, and residual feature value in the sitting posture feature value array is as follows: A linear fit is performed on all data points in the trend term of the sight distance data within the left neighbor time range using the least squares method, and the slope of the linear fit result is used as the trend feature value; a Fourier transform is performed on the periodic term of the sight distance data within the left neighbor time range to obtain the corresponding spectrum data, and the spectrum data is used as the input of the principal component analysis algorithm to obtain the eigenvector corresponding to the maximum eigenvalue obtained during the processing of the spectrum data by the principal component analysis algorithm, which is used as the periodic feature value; the average value of all data points in the residual term of the sight distance data within the left neighbor time range is used as the residual feature value.

[0077] It should be noted that, due to the influence of personal habits, users often work and study in their usual sitting posture when adjusting their posture. Each adjustment is influenced by the previous posture, forming a behavioral pattern. This invention selects to obtain corresponding feature values ​​from the trend, periodic, and residual terms of the viewing distance data within the left neighbor time range to form a corresponding posture feature value array to describe the behavioral pattern information contained in short-term posture changes. Specifically, the trend feature value reflects the directional tendency of posture adjustment, such as continuous forward or backward leaning; the periodic feature value captures the periodic pattern of posture adjustment, such as high-frequency small-amplitude swaying or low-frequency large-amplitude changes; and the residual feature value characterizes the intensity of random disturbances, such as sudden posture changes. Therefore, the posture feature value array does not simply describe the viewing distance data itself, but highly condenses the behavioral pattern information of user posture adjustment, reflecting the user's adjustment inertia influenced by habits in the short term, such as the behavioral inertia of repeatedly fine-tuning the viewing distance to maintain a comfortable posture during work and study.

[0078] Step S303: Cluster all corresponding times based on the sitting posture feature value array. Calculate the user's sitting posture adjustment rate at the current time based on the relative position of the current time in its respective cluster and the distribution of all times in the cluster.

[0079] As a preferred embodiment, the method for obtaining the user's posture adjustment rate at the current moment is as follows: using the Euclidean distance between the posture feature value arrays at different times as a distance metric, and performing clustering processing on all times using the DBSCAN clustering algorithm to obtain several clusters, denoting any cluster as a posture time cluster, and analyzing the user's posture adjustment rate at the current moment by using the distribution of all times in the posture time cluster in the cluster space and the relative position of the current moment in the cluster space.

[0080] It should be noted that, since the sitting posture feature value array is obtained based on the duration of the left neighbor in this embodiment of the invention, and the specific size of the duration of the left neighbor is adjusted by the sitting posture variation coefficient, the value of the feature value element in the sitting posture feature value array is highly correlated with the user's sitting posture adjustment behavior. Therefore, when clustering based on the sitting posture feature value array, moments with similar behavior patterns are grouped into the same cluster, so that each sitting posture moment cluster corresponds to a typical sitting posture adjustment behavior pattern. For example, in the scenario of subtle posture correction during focused reading, the corresponding sitting posture moment cluster may represent a high-frequency small-amplitude adjustment pattern; while in the scenario of significant posture reset after long-term work, the corresponding sitting posture moment cluster may represent a low-frequency large-amplitude adjustment pattern. In any cluster of sitting posture moments, the closer the value is to the center of the cluster (i.e., the cluster center), the more closely the viewing distance change characteristics exhibited at that moment within its corresponding left neighbor time range match the behavioral pattern of user posture adjustment represented by that cluster. In other words, in the clustering results, the cluster center of each sitting posture moment cluster represents the ideal prototype of that behavioral pattern. For example, in a cluster of sitting posture moments corresponding to high-frequency, small-amplitude adjustments, the feature value array corresponding to the center point describes the most typical adjustment rhythm, such as a stable periodic fine-tuning slope. The distance between the feature value at any moment within the cluster and the center directly reflects the degree of conformity between the behavior and the pattern. The closer the value is to the center, the more closely the viewing distance change characteristics (such as trend slope and spectral dominance) exhibited at that moment within its left neighbor time range match the behavioral pattern defined by that cluster.

[0081] Furthermore, it should be noted that the sitting posture time clusters obtained in the embodiments of the present invention become a discretized and semantic carrier of the user's sitting posture adjustment behavior pattern. Each cluster corresponds to a recognizable behavioral fingerprint. It extracts behavioral features through a dynamic window of the left neighborhood and uses clustering to map high-dimensional features into low-dimensional behavioral patterns. This not only avoids the mechanical nature of traditional time segmentation, but also realizes the leap from data points of viewing distance data to behavioral patterns, making sitting posture adjustment behavior quantifiable and comparable. Ultimately, it can be used for subsequent accurate sitting posture and viewing distance monitoring. Its underlying logic is consistent with the inertial nature of human sitting posture adjustment, that is, behavioral patterns originate from short-term historical influences, and clustering provides a structured expression for this influence.

[0082] As an optional embodiment, the method for obtaining the posture adjustment rate at the current moment is as follows: if the current moment is an outlier in the cluster space, the posture adjustment rate at the current moment is set to 0; if the current moment is not an outlier in the cluster space, the posture moment cluster to which the current moment belongs is obtained, referred to as the current posture moment cluster; the standard deviation of the corresponding distance between all moments in the current posture moment cluster and the cluster center is obtained; the average distance between all moments in the current posture moment cluster and the cluster center is recorded as the average distance parameter of the current posture moment cluster; the corresponding distance between the current moment and the cluster center of the current posture moment cluster is recorded as the distance parameter of the current moment; the user's posture adjustment rate at the current moment is calculated based on the standard deviation of the corresponding distance between all moments in the current posture moment cluster and the cluster center, and the difference between the average distance parameter and the distance parameter of the current moment.

[0083] As an optional embodiment, the specific calculation method for the user's posture adjustment rate at the current moment is as follows: ;in, This indicates the user's posture adjustment rate at the current moment; This represents the standard deviation of the distances between all times in the current sitting posture time cluster and the cluster center; This represents the average distance between all times in the current sitting posture time cluster and the cluster center; This represents the distance between the current time and the cluster center in the current sitting posture time cluster; This represents the sigmoid normalization function.

[0084] It should be noted that the posture adjustment rate describes the probability that a user needs to adjust their posture at the current moment due to the influence of their short-term sitting posture. A higher posture adjustment rate indicates a greater probability. The calculation method for the posture adjustment rate is based on DBSCAN cluster analysis to analyze short-term sitting behavior patterns and predict whether a user is about to adjust their posture. The distance difference is used in this calculation. This reflects the relative position of the current moment within the current sitting posture time cluster. A larger value indicates that the current moment's relative position within the current sitting posture time cluster is closer to the cluster center, thus more closely reflecting the sitting posture adjustment behavior pattern corresponding to the current sitting posture time cluster. Furthermore, dividing by... This is to eliminate the influence of cluster size.

[0085] Step S304: Combine the size of the adjustment cycle to which the current moment belongs, the posture change coefficients of other adjustment cycles before the current adjustment cycle, and the user's posture adjustment rate at the current moment to construct a polling coefficient, thereby adjusting the initial polling frequency and obtaining the polling frequency of the smart card reader at the current moment.

[0086] As an optional embodiment, the specific calculation method for the polling frequency at the current moment is as follows: ;

[0087] in, This indicates the polling frequency of the reader at the current moment; Indicates the preset initial polling frequency; Indicates the preset adjustment ratio; This indicates the specific value of the adjustment cycle to which the current moment belongs; This indicates the user's posture adjustment rate at the current moment; This represents the average of the sitting posture variation coefficients from other adjustment cycles preceding the adjustment cycle.

[0088] It should be noted that by dynamically adjusting the polling frequency of the reader in the smart card, it is possible to effectively acquire the user's line-of-sight data and avoid data redundancy, while also reducing the energy consumption of the reader in the smart card, thereby extending the battery life of the smart card. In addition, in this embodiment of the invention, the adjustment multiplier is preset to 2 based on experience, which can be adjusted according to the actual situation. This embodiment of the invention does not impose specific limitations.

[0089] Thus, the polling frequency at the current moment is obtained through the above method.

[0090] Step S004: The reader collects the line-of-sight data at the current moment using the current adjustment cycle and polling frequency.

[0091] Specifically, firstly, the reader's operating parameters are dynamically configured based on the current adjustment cycle and polling frequency.

[0092] As a preferred embodiment, the specific method for dynamically configuring the reader's operating parameters is as follows: the reader is set to continuously execute a complete cyclic process of "activation—sending command—receiving response—ranging calculation—Bluetooth upload" at the current polling frequency within the current adjustment cycle, with each cycle interval being [time missing]. Second.

[0093] Then, at the end of each adjustment cycle, the system automatically caches all line-of-sight data collected within that cycle and records it as the [number]th [cycle]. The first adjustment cycle dataset is used for subsequent updates to the historical regularity coefficient, sitting posture variation coefficient, and cycle adjustment coefficient calculation, serving as the basis for generating the first adjustment cycle dataset. The input basis for each adjustment cycle.

[0094] Combining the above steps, and so on, each time the number is obtained... After considering the polling frequency within the first adjustment cycle and the line-of-sight data obtained in that adjustment cycle, the first... The sight distance data within the adjustment cycle is obtained. The size of each adjustment cycle.

[0095] The online learning capability achieved through the embodiments of the present invention can automatically adjust the monitoring intensity according to the user's evolving habits without manual intervention. This ensures high-precision capture through high-frequency polling when sitting posture changes frequently, while significantly reducing energy consumption through low-frequency polling during stable eye use, thus achieving optimal synergy between energy consumption and accuracy.

[0096] Thus, the operating parameters of the reader / writer are obtained through the above method.

[0097] Step S005: Set a safe viewing distance threshold based on user information, and monitor sitting posture based on the viewing distance value at the current moment in the viewing distance data.

[0098] Specifically, firstly, a baseline safe distance, height compensation parameters, and age compensation parameters are preset and combined with user information to calculate the user's safe viewing distance threshold.

[0099] As an optional embodiment, the specific calculation method for the safe line-of-sight threshold is as follows: ;in, This indicates the user's safe line-of-sight threshold. This is the preset baseline safety distance; User's height (unit: m); and These represent the preset height compensation parameters and age compensation parameters, respectively. User's age.

[0100] It should be noted that in this embodiment of the invention, the preset baseline safety distance is 30cm based on experience. In other embodiments, it can be adjusted according to the "Guidelines for Appropriate Technologies for the Prevention and Control of Myopia in Children and Adolescents". Furthermore, regarding the height compensation parameter and age compensation parameter, this embodiment of the invention uses the height compensation parameter to compensate for visual angle deviations caused by height differences, and the age compensation parameter to gradually tighten the safety threshold according to the age growth trend, so as to provide more stringent protection for the visual development period of adolescents. Specifically, the height compensation parameter is preset in this embodiment of the invention. Age compensation parameters The values ​​can be adjusted according to the actual situation. The specific values ​​of the benchmark safety distance, height compensation parameter and age compensation parameter are not specifically limited in the embodiments of the present invention.

[0101] Then, if the viewing distance value in the viewing distance data at the current moment is less than the user's safe viewing distance threshold, an alarm message will be sent through the mobile APP to notify the user to adjust their sitting posture.

[0102] By following the steps above, dynamic monitoring of the user's posture and viewing distance can be achieved using RFID tag identification technology.

[0103] Please see Figure 3 The diagram shows a structural block diagram of a dynamic monitoring system for sitting posture and line of sight based on radio frequency tag identification provided in an embodiment of the present invention. The system includes a memory 302, a processor 301, and a computer program 3021 stored in the memory 302 and executable on the processor. When the processor 301 executes the computer program 3021, it implements steps S001 to S005 of the dynamic monitoring method for sitting posture and line of sight based on radio frequency tag identification.

[0104] Furthermore, in an optional embodiment, the memory 302 described above may include read-only memory and random access memory, and provide instructions and data to the processor. The memory 302 may also include non-volatile random access memory. For example, the memory may also store device type information.

[0105] It should be noted that the embodiments used in this example The model is only used to represent negative correlations and the results of the constraint model output are in Within this range, in specific implementations, other models with the same purpose can be substituted; this embodiment is merely an example. The description will be based on a model, without making specific limitations on it. This refers to the input of the model.

[0106] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for dynamic monitoring of sitting posture and visual distance based on radio frequency tag identification, characterized in that, The method comprises the following steps: Obtain user information, preset an initial polling frequency and an initial collection time length, and collect line-of-sight data through a reader; Decompose the line-of-sight data, analyze the user's sitting posture adjustment characteristics by using the data decomposition result, adjust the initial collection time length, and determine the adjustment period at the current time, including the following specific method: The line-of-sight data collected by the reader within the time range corresponding to the initial collection time length after the reader is started and at the initial polling frequency is taken as the preliminary line-of-sight data; Decompose and process the preliminary line-of-sight data by using an STL time sequence decomposition algorithm to obtain a trend item, a periodic item and a residual item of the preliminary line-of-sight data, calculate a historical rule coefficient of the user in the preliminary line-of-sight data according to the data distribution in the trend item, the periodic item and the residual item; The historical law coefficient is used to adjust the initial collection time length to obtain a specific value of a next adjustment period. The specific value of the next adjustment period is calculated by: wherein represents the specific value of the i-th adjustment period; represents the preset adjustment multiple; represents the historical law coefficient of the user in the line-of-sight data in the i-th adjustment period; and represents the specific value of the i-th adjustment period. represents the preset adjustment multiple. represents the historical law coefficient of the user in the line-of-sight data in the i-th adjustment period. represents the specific value of the i-th adjustment period. The adjustment period refers to the collection time length of the reader when collecting the line-of-sight data; Cluster all time points in the line-of-sight data based on the fluctuation range of the line-of-sight data and the distribution characteristics of the data points in the decomposition result, adjust the initial polling frequency according to the data distribution of the clustering result and the relative position of the current time, and obtain the polling frequency at the current time; Make the reader collect the line-of-sight data at the current time by using the adjustment period and the polling frequency at the current time; Set a safe line-of-sight threshold in combination with the user information, and perform sitting posture monitoring according to the line-of-sight value at the current time in the line-of-sight data. 2.The method of claim 1, wherein, The specific method for obtaining the historical rule coefficient is as follows: Obtain the standard deviation of the slope of all data points in the seasonal item of the preliminary line-of-sight data, perform fast Fourier transform on the periodic item of the preliminary line-of-sight data to obtain a frequency spectrum diagram of the periodic item, obtain the number of frequencies with a non-zero amplitude in the frequency spectrum diagram, calculate the amplitude cumulative value of all frequencies with a non-zero amplitude in the frequency spectrum diagram, and obtain the average value of all data points in the residual item of the preliminary line-of-sight data; and obtain the historical rule coefficient according to the standard deviation of the slope, the number of frequencies with a non-zero amplitude in the frequency spectrum diagram, the amplitude cumulative value and the average value of all data points in the residual item. 3.The method of claim 1, wherein, The determining of the adjustment period of the current time further includes: after the specific value of the first adjustment period is obtained , when at least 2 adjustment periods have passed before the first adjustment period, calculating a period adjustment coefficient of the first adjustment period according to the average value and the standard deviation of the specific values corresponding to all the adjustment periods before the first adjustment period, multiplying the specific value of the first adjustment period by the period adjustment coefficient to obtain the specific value of the adjusted first adjustment period. ​​​​​ 4.The method of claim 1, wherein, The specific method for obtaining the polling frequency at the current time by clustering all time points in the line-of-sight data based on the fluctuation range of the line-of-sight data and the distribution characteristics of the data points in the decomposition result, adjusting the initial polling frequency according to the data distribution of the clustering result and the relative position of the current time, is as follows: Analyze the sitting posture change coefficient of the user in the initial collection time length range corresponding to the preliminary line-of-sight data by using the discrete characteristics and fluctuation degree shown in the trend item, the periodic item and the residual item of the preliminary line-of-sight data; Preset a basic left neighborhood time length, obtain the left neighborhood time length of the adjustment period in combination with the sitting posture change coefficient of the adjustment period, and extract a sitting posture characteristic value array formed by a plurality of characteristic values according to the change of the line-of-sight data within the corresponding left neighborhood time length at any time of the adjustment period. The sitting posture characteristic value array is used to cluster all corresponding time points, and a sitting posture adjustment rate of the user at the current time point is calculated according to a relative position of the current time point in a cluster to which the current time point belongs and a distribution of all time points in the cluster. 5.The method of claim 4, wherein, The specific method for obtaining the sitting posture variation coefficient is as follows: An average value of information entropy of all STL time sequence decomposition results of the pre-stage sight distance data and a range of data points in the pre-stage sight distance data are obtained, and the sitting posture variation coefficient is obtained according to the average value of information entropy and the range of data points in the pre-stage sight distance data, wherein the average value of information entropy and the range of data points in the pre-stage sight distance data are positively correlated with the sitting posture variation coefficient. 6.The method of claim 4, wherein, The specific method for obtaining the sitting posture characteristic value array is as follows: Pre-set left-neighborhood time length of basis According to the method for obtaining the sitting posture variation coefficient, the sitting posture variation coefficient of the user under any adjustment period is calculated, and the data points in the trend item, the period item and the residual item of the range of the line-of-sight data before any time and belonging to the corresponding left-neighborhood time length of the adjustment period to which the time belongs are obtained. The trend feature value, the period feature value and the residual feature value at the time are calculated by using the distribution of the data points in the trend item, the period item and the residual item of the line-of-sight data in the range of the left-neighborhood time length, and the array formed by the trend feature value, the period feature value and the residual feature value at the time is taken as the sitting posture feature value array of the time. 7.The method of claim 4, wherein, The specific method for clustering all corresponding time points based on the sitting posture characteristic value array is as follows: Euclidean distances between the sitting posture characteristic value arrays at different time points are used as distance measurement methods, and all time points are clustered by using a DBSCAN clustering algorithm to obtain a plurality of clusters, and any cluster is referred to as a sitting posture time point cluster. 8.The method of claim 4, wherein, The specific method for obtaining the sitting posture adjustment rate is as follows: If the current time point is an outlier in the cluster space, the sitting posture adjustment rate at the current time point is set to 0; if the current time point is not an outlier in the cluster space, a sitting posture time point cluster in which the current time point is located is obtained and is referred to as a current sitting posture time point cluster, a standard deviation of distances between all time points in the current sitting posture time point cluster and a cluster center is obtained, an average value of distances between all time points in the current sitting posture time point cluster and the cluster center is referred to as an average distance parameter of the current sitting posture time point cluster, a distance between the current time point and the cluster center of the current sitting posture time point cluster is referred to as a distance parameter of the current time point, and a sitting posture adjustment rate of the user at the current time point is calculated according to the standard deviation of distances between all time points in the current sitting posture time point cluster and the cluster center, and a difference between the average distance parameter and the distance parameter of the current time point. 9.A system for dynamic monitoring of sitting posture and line-of-sight based on radio frequency tag identification, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor implements the steps of the method for dynamically monitoring a sitting posture and a sight distance based on radio frequency tag identification according to any one of claims 1-8 when the processor executes the computer program.

Images