Intelligent sports data analysis management system and method based on Internet of Things

Abstract

The invention discloses an intelligent sports data analysis and management system and method based on the Internet of Things, and particularly relates to the technical field of data transmission. The method comprises the following steps: selecting RFID tags, constructing a multi-tag intensive activity test site, simulating a conflict environment of a high-density scene, recording a signal intensity value and analyzing a fluctuation amplitude; when the signal intensity fluctuation amplitude is large, the signal propagation time value is further recorded, the deviation condition is judged, and the accuracy of decoding the data information of the multiple labels by the card reader is comprehensively evaluated by combining the signal propagation time value and the deviation condition; according to the evaluation result, the decoding result is divided into three classes, and a classification management method is adopted; for an incompletely accurate decoding result, interpolation calculation is carried out by using positions and time of front and back known points, and a kinematics model is dynamically adjusted in combination with a motion mode of an athlete so as to optimize the accuracy of the decoding result, so that the problem of decoding errors caused by label collision in the prior art is effectively solved; and the continuity and the reliability of data acquisition are obviously improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data transmission, and in particular to a smart sports data analysis management system and method based on the Internet of Things. Background Art

[0002] Smart sports data analysis and management based on the Internet of Things refers to connecting sensors, smart devices and communication networks through the Internet of Things (IoT) technology to achieve real-time collection, transmission, analysis and management of sports data. Specifically, this system can obtain physiological data (such as heart rate, body temperature) and motion data (such as speed, displacement) from athletes' wearable devices (such as smart bracelets, smart shoes) or sensor devices in the venue, and analyze and process the data through cloud or edge computing technology to provide personalized training suggestions, event data statistics, venue management optimization and other functions. This model not only improves the efficiency of sports training and event management, but also provides strong support for scientific sports decision-making.

[0003] In the existing technology, the application of IoT in the field of sports has been quite extensive. For example, the athlete location tracking system based on RFID (radio frequency identification) and GPS technology can monitor the athlete's movement trajectory in real time; smart wearable devices can collect sports data and transmit it to mobile devices or the cloud via Bluetooth or Wi-Fi; the analysis platform combining big data and artificial intelligence can deeply mine the value of sports data for evaluating athlete performance or predicting game results.

[0004] The prior art has the following deficiencies:

[0005] In large-scale, dense scenarios (such as marathons or group competitions), if multiple RFID tags simultaneously backscatter signals, the antenna may receive mixed signals, resulting in tag collisions, which will cause the reader to be unable to correctly distinguish and decode the data information of multiple tags. And because of signal conflicts, some tag data decoding errors may occur, and the athlete's trajectory may jump, be discontinuous, or even completely wrong. For example, the system may record that the athlete suddenly deviates from the track at a certain point in time. In addition, incorrect trajectories will mislead coaches and athletes in their analysis of performance, affect sports plans and technical action adjustments; in competitions, it may also cause controversy and question the fairness and authority of the data monitoring system. Summary of the invention

[0006] The purpose of the present invention is to provide a smart sports data analysis management system and method based on the Internet of Things to solve the shortcomings of the background technology.

[0007] In order to achieve the above object, the present invention provides the following technical solution: a smart sports data analysis and management method based on the Internet of Things, comprising the following steps:

[0008] S1: Select several RFID tags with unique IDs, build a test site with intensive multi-tag activities, and set the number of RFID tags and the range of movement speed in the test site;

[0009] S2: In the reader coverage area, gradually increase the number of tags and control the distribution density of tags to simulate the conflict environment of high-density scenarios. Under different numbers and densities of tags, record the signal strength value of each tag and analyze the fluctuation range of the signal strength value.

[0010] S3: If the fluctuation range of the signal strength value is large, the signal propagation time value of each RFID tag is recorded, and the deviation of the signal propagation time value in the same time window is determined;

[0011] S4: Evaluate the accuracy of data information decoded by the card reader for multiple tags based on the fluctuation amplitude of the signal strength value of the RFID tag and the deviation of the signal propagation time value in the same time window;

[0012] S5: according to the evaluation result of the accuracy of the data information of the multiple tags decoded by the card reader, the decoding results of the card reader are divided into accurate decoding results, incompletely accurate decoding results and inaccurate decoding results, and corresponding management is performed;

[0013] S6: For decoding results with incomplete accuracy, the positions and times of the known points of the two RFID tags before and after are used for interpolation calculations, and the kinematic model is dynamically adjusted according to the calculation results and the athlete's movement pattern to optimize the accuracy of the data decoding results.

[0014] Preferably, in S2, the signal strength value of each tag is recorded, and the fluctuation amplitude of RSSI data with different distribution densities is analyzed to generate an RSSI density distribution fluctuation index. The RSSI density distribution fluctuation index is obtained by:

[0015] Collect RSSI signal strength value x(s) and record its sequence changing over time ; Sample the RSSI signal at a fixed sampling frequency to obtain equally spaced discrete time point data. N is the total length of the RSSI signal sequence. Select the window function w(s), set the window length L, and divide the signal strength value x(s) into multiple segments on the time axis according to the window function w(s). Each segment is L in length and weighted. For each segment, calculate its Fourier transform to obtain the frequency component of the signal. The expression is: ; where the window center is at τ, is the Fourier transform result of the signal strength value x(s), is the kernel function of Fourier transform, is the frequency. In discrete time signals, STFT is expressed as: ; Calculate the frequency components of each window Amplitude value , and record it, τ is the window center position on the time axis, is the discrete frequency component, representing the discrete points on the frequency axis, n is the number of windows, is the kernel function of the discrete transformation, recording all time windows τ and frequency components The amplitude , and obtain the time-frequency distribution matrix , calculate the total energy in each time window, the expression is: ; Represents the frequency energy density of the signal at time τ, counts the mean μE and standard deviation σE of energy fluctuations in all time windows, and calculates the RSSI density distribution fluctuation index, which is expressed as: ; In the formula, It is the RSSI density distribution fluctuation index.

[0016] Preferably, in S3, after analyzing the deviation of the signal propagation time value in the same time window, a TOF propagation deviation index is generated, and the TOF propagation deviation index is obtained by:

[0017] Collect a sequence of TOF values ​​within the same time window ; Where M is the total number of RFID tags in the current time window, each TOF value represents the signal propagation time of a tag, and the TOF value sequence is sorted from small to large, recorded as ; Calculate the quartiles: the first quartile Q1 is the TOF value in the lower quartile after sorting, the third quartile Q3 is the TOF value in the upper quartile after sorting, and the interquartile range IQR: IQR=Q3−Q1; the upper and lower limits are set as: Lower limit: ; Upper bound: ; The TOF values ​​that are less than the lower bound and greater than the upper bound are recorded as outliers, and the proportion of outliers to the total TOF values ​​in the same time window is calculated, that is, the TOF propagation deviation index is calculated.

[0018] Preferably, in S4, the accuracy of data information decoded by the card reader for multiple tags is evaluated according to the fluctuation amplitude of the signal strength value of the RFID tag and the deviation of the signal propagation time value in the same time window, specifically:

[0019] The TOF propagation deviation index and the RSSI density distribution fluctuation index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the accuracy value label of the data information decoded by the card reader for multiple tags as the prediction target, and takes minimizing the sum of the prediction errors of the accuracy value labels of the data information decoded by all readers for multiple tags as the training target. The machine learning model is trained until the sum of the prediction errors converges, and the model training is stopped. The accuracy value of the data information decoded by the card reader for multiple tags is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0020] Preferably, in S5, according to the evaluation result of the accuracy of the data information of the multiple tags decoded by the card reader, the decoding result of the card reader is divided into an accurate decoding result, an incompletely accurate decoding result and an inaccurate decoding result, specifically:

[0021] Compare the acquired accuracy value of the data information of the multiple tags decoded by the card reader with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and compare the accuracy value of the data information of the multiple tags decoded by the card reader with the first standard threshold and the second standard threshold respectively;

[0022] If the accuracy value of the data information decoded by the card reader for multiple tags is greater than the second standard threshold, it means that the accuracy of the data information decoded by the card reader for multiple tags is high, and a high-accuracy decoding signal is generated at this time, and the decoding result of the card reader is classified as an accurate decoding result;

[0023] If the accuracy value of the data information decoded by the card reader for multiple tags is greater than or equal to the first standard threshold value and less than or equal to the second standard threshold value, it means that the accuracy of the data information decoded by the card reader for multiple tags is average, and a medium accuracy decoding signal is generated at this time, and the decoding result of the card reader is classified as an incomplete accuracy decoding result;

[0024] If the accuracy value of the data information decoded by the card reader for multiple tags is less than the first standard threshold, it means that the accuracy of the data information decoded by the card reader for multiple tags is low. At this time, a low-accuracy decoding signal is generated, and the decoding result of the card reader is classified as an inaccurate decoding result.

[0025] Preferably, in S6, for the incomplete accuracy decoding result, the positions and times of the known points of the two preceding and following RFID tags are used for interpolation calculation, and the kinematic model is dynamically adjusted according to the calculation result and the movement pattern of the athlete, specifically:

[0026] For the decoding result with incomplete accuracy, the position and time of the two known points before and after are used to calculate the position of the current time point. The calculation expression is: ; In the formula, is the predicted position corresponding to the current time point t, is the location of the previous known point of the RFID tag, After the RFID tag knows the location of the point, is the timestamp of the previous known point, is the timestamp of the next known point, t is the timestamp of the location point to be predicted, satisfying < < ;

[0027] When the athlete's movement pattern changes, the interpolation calculation needs to be dynamically adjusted in combination with the kinematic model. The uniform motion model expression is: ; where v is the speed of the athlete, ; The calculation expression of uniform acceleration motion model is: ; In the formula, a is the athlete's acceleration, which is calculated based on the speed change between the two known points before and after. ; is the velocity at the previous moment, ; The calculation expression of the curve motion model is: ; In the formula, , , where is the coordinate of the center of the arc, r is the radius of the curve, which is determined by the geometric relationship between the two positions before and after. is the center angle of the athlete at the current time point, ω is the angular velocity of the athlete, is the predicted position at the current time point;

[0028] Combined with the motion mode, interpolation calculation is performed according to dynamic conditions, and the comprehensive formula is expressed as: ;

[0029] The current position is directly calculated through the two known points before and after, and the uniform speed, acceleration and curve movement modes are judged according to the athlete's speed change and trajectory characteristics; the kinematic model is selected based on the movement mode, the interpolation formula is adjusted, the decoding result is optimized, and the optimized decoding result is compared with the adjacent known points to judge the optimization effect. If the optimization effect is good, it will be stored in the database.

[0030] Preferably, the optimized decoding result is compared with the adjacent known points to determine the optimization effect, specifically:

[0031] Calculate the time deviation between the optimized decoding result and the adjacent known points. The expression is: ; In the formula, is the time point of the optimized decoding result, is a time point adjacent to a known point;

[0032] Calculate the distance deviation between the optimized decoding result and the adjacent known points in space: ; is the position coordinate of the optimized decoding result, is the position coordinates of the nearby known points;

[0033] The optimization effect can be judged by comparing the trajectory continuity of the optimized decoding results with the known points before and after: ; In the formula, To optimize the distance between the result and the known points before and after, is the time interval between the previous and next known points, Should be less than the time threshold Tthresh, Δd should be less than the spatial error threshold Dthresh, if the trajectory continuity C is within the preset standard range of the athlete's actual movement speed , the optimization effect is considered good.

[0034] The present invention also provides a smart sports data analysis and management system based on the Internet of Things, including a test environment construction module, a signal strength acquisition and analysis module, a signal propagation time analysis module, a decoding accuracy evaluation module, a decoding result classification management module and a decoding result optimization module:

[0035] Test environment construction module: select several RFID tags with unique IDs, build a test site with intensive multi-tag activity, and set the number of RFID tags and the range of movement speed in the test site;

[0036] Signal strength collection and analysis module: In the reader coverage area, gradually increase the number of tags and control the distribution density of tags to simulate the conflict environment of high-density scenes. Under different numbers and densities of tags, record the signal strength value of each tag and analyze the fluctuation range of the signal strength value.

[0037] Signal propagation time analysis module: If the fluctuation range of the signal strength value is large, the signal propagation time value of each RFID tag is recorded, and the deviation of the signal propagation time value in the same time window is determined;

[0038] Decoding accuracy evaluation module: evaluates the accuracy of data information decoded by the card reader for multiple tags based on the fluctuation of the signal strength value of the RFID tag and the deviation of the signal propagation time value in the same time window;

[0039] Decoding result classification management module: according to the evaluation results of the accuracy of the data information of multiple tags decoded by the card reader, the decoding results of the card reader are divided into accurate decoding results, incompletely accurate decoding results and inaccurate decoding results, and corresponding management is performed;

[0040] Decoding result optimization module: For decoding results with incomplete accuracy, the positions and times of the known points of the two front and rear RFID tags are used for interpolation calculations, and the kinematic model is dynamically adjusted according to the calculation results and the athlete's movement pattern to optimize the accuracy of the data decoding results.

[0041] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0042] 1. The present invention combines RFID technology, signal processing algorithms and kinematic models to achieve efficient decoding and accurate data analysis of RFID tags in large-scale dense scenarios, effectively solving the problems of decoding errors, motion trajectory jumps and data discontinuity caused by tag collisions in the prior art. By recording and analyzing the signal strength value (RSSI) and the signal propagation time value (TOF), the density distribution fluctuation index and the propagation deviation index are generated, and the decoding accuracy is evaluated using a polynomial regression model. The decoding strategy and optimized data are dynamically adjusted according to the results, ensuring the high reliability and scientificity of the decoding results. The present invention innovatively combines interpolation calculations and motion pattern recognition to dynamically adjust the kinematic model, correct and optimize the incompletely accurate decoding results, and further improve the continuity and rationality of data decoding.

[0043] 2. The present invention significantly improves the decoding performance of RFID signals in large-scale dense environments, ensures the accuracy of athlete sports data collection, and avoids the negative impact of erroneous trajectories on sports analysis, game decisions, and fairness. By classifying and dynamically optimizing the decoding results, the real-time and robustness of smart sports data analysis are enhanced, making it suitable for complex scenarios such as marathons and collective competitions, promoting the in-depth application of Internet of Things technology in smart sports, and providing strong support for athlete training improvements and scientific event organization. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0045] Figure 1 The figure is a flow chart of the method of the present invention.

[0046] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] Example 1, please refer to Figure 1 and Figure 2 As shown, the smart sports data analysis and management method based on the Internet of Things described in this embodiment includes the following steps:

[0049] S1: Select several RFID tags with unique IDs, build a test site with intensive multi-tag activities, and set the number of RFID tags and the range of movement speed in the test site;

[0050] S2: In the reader coverage area, gradually increase the number of tags and control the distribution density of tags to simulate the conflict environment of high-density scenarios. Under different numbers and densities of tags, record the signal strength value of each tag and analyze the fluctuation range of the signal strength value.

[0051] S3: If the fluctuation range of the signal strength value is large, the signal propagation time value of each RFID tag is recorded, and the deviation of the signal propagation time value in the same time window is determined;

[0052] S4: Evaluate the accuracy of data information decoded by the card reader for multiple tags based on the fluctuation amplitude of the signal strength value of the RFID tag and the deviation of the signal propagation time value in the same time window;

[0053] S5: according to the evaluation result of the accuracy of the data information of the multiple tags decoded by the card reader, the decoding results of the card reader are divided into accurate decoding results, incompletely accurate decoding results and inaccurate decoding results, and corresponding management is performed;

[0054] S6: For decoding results with incomplete accuracy, the positions and times of the known points of the two RFID tags before and after are used for interpolation calculations, and the kinematic model is dynamically adjusted according to the calculation results and the athlete's movement pattern to optimize the accuracy of the data decoding results.

[0055] In S1, several RFID tags with unique IDs are selected to construct a test field with intensive multi-tag activities, and the number and movement speed range of RFID tags are set in the test field, specifically:

[0056] Select several RFID tags with unique IDs. Passive or active tags can be used to ensure the diversity of tag types according to the experimental requirements. Initialize the configuration of each tag to ensure that the unique identification (ID) of each tag can be correctly identified and recorded. Arrange the test site in the selected area. The site design should simulate the real scene of intensive multi-tag activities, such as a racetrack, indoor venue or training area. Configure RFID readers to cover key locations of the test site (such as starting points, curves, and sprint finishes) to ensure that the card reading range matches the tag activity area. Set the number of tags used in the experiment (such as 10, 20, 50), and gradually increase the tag density to simulate a high-density conflict environment. Control the movement speed range of the tag to cover static state, low-speed movement (such as jogging) and high-speed sprinting (such as sprinting) to reflect the complexity of different sports scenes.

[0057] S2: In the reader coverage area, gradually increase the number of tags and control the distribution density of tags to simulate the conflict environment of high-density scenarios. Under different numbers and densities of tags, record the signal strength value of each tag and analyze the fluctuation range of the signal strength value.

[0058] Arrange one or more RFID readers in the experimental area to cover the entire test site. Make sure the reader supports the recording function of the signal strength change characteristic (RSSI) and is connected to the data processing system for real-time data recording and analysis.

[0059] Prepare several RFID tags with unique IDs (e.g., 10, 20, 50) and add them to the site in batches. Configure a unique ID for each tag and ensure that its signal can be recognized by the reader. Increase the number of tags gradually from small to large, simulating a multi-tag environment from low density to high density. Adjust the spacing between tags from uniform distribution to random dense distribution, and observe the changes in signal characteristics under different density conditions.

[0060] Create a real tag conflict environment and record the signal strength value in the case of multiple tags. Initially place 10 tags and record their signal strength values ​​as baseline data; then gradually increase the number of tags (such as 10 tags each time) until the maximum tag density of the experiment is reached (such as 50 or more). Adjust the distribution density, including: Uniform distribution: Arrange the tags at a fixed spacing and record the stability of the RSSI value. Random distribution: Arrange the tags randomly to simulate the irregular and dense environment in the real scene and record the RSSI fluctuation range. Under each tag number and density condition, record the RSSI values ​​of all tags, generate a time series data table, and mark the signal fluctuation range and outliers. If some tags are not recognized by the reader, record their loss.

[0061] For each tag, within a fixed period of time (such as 1 minute), record the maximum, minimum and average RSSI values: ; Group and count RSSI data by number of tags and distribution density, and calculate the average fluctuation amplitude under different conditions. Identify tags with abnormal signal strength fluctuations (such as tags with fluctuation amplitudes outside the expected range) and mark possible conflicts or interference.

[0062] After analyzing the fluctuation amplitude of RSSI data with different distribution densities, the RSSI density distribution fluctuation index is generated. The RSSI density distribution fluctuation index is obtained as follows:

[0063] Collect RSSI signal strength value x(t) and record its sequence changing over time ; Sample the RSSI signal at a fixed sampling frequency to obtain equally spaced discrete time point data, where N is the total length of the RSSI signal sequence (number of sampling points). Select a window function w(t) (such as a rectangular window, Hanning window, or Gaussian window) to intercept a portion of the signal for local frequency analysis. Set the window length L, which should be adjusted according to the time resolution and frequency resolution requirements of the signal.

[0064] On the time axis, the signal strength value x(s) is divided into multiple signal segments according to the window function w(s), each segment is L in length, and weighted. For each segment, its Fourier transform (FT) is calculated to obtain the frequency component of the signal. The expression is: ; where the window center is at τ, is the Fourier transform result of the signal strength value x(s), is the kernel function of Fourier transform, is the frequency. In discrete time signals, STFT is expressed as: ; Calculate the frequency components of each window Amplitude value , and record it, τ is the window center position on the time axis (each moment corresponds to a window), is the discrete frequency component, representing the discrete points on the frequency axis (k=0,1,…,L−1), n ​​is the number of windows, is the kernel function of the discrete transformation, recording all time windows τ and frequency components The amplitude , and obtain the time-frequency distribution matrix . Calculate the total energy in each time window, the expression is: ; Represents the frequency energy density of the signal at time τ. Statistically calculate the mean μE and standard deviation σE of the energy fluctuations in all time windows, and calculate the RSSI density distribution fluctuation index, the expression is: ; In the formula, It is the RSSI density distribution fluctuation index.

[0065] When the RSSI density distribution fluctuation index is large, it means that the distribution fluctuation of signal strength in time and space is more obvious. This usually means that there is a strong conflict or interference between the signals of multiple tags, which makes it more difficult for the card reader to distinguish and decode tag data. Therefore, the larger the RSSI fluctuation index, the lower the decoding accuracy of the card reader may be, especially in a high-density tag environment, which is prone to missed reading or misreading, affecting the integrity and reliability of the data.

[0066] When the RSSI density distribution fluctuation index is small, it means that the distribution of signal strength is more stable and the fluctuation amplitude is small. This indicates that there is less interference with the tag signal, and the card reader can more easily distinguish the data signals of each tag, thereby improving the success rate and accuracy of decoding. Therefore, the smaller the RSSI fluctuation index, the better the decoding performance of the card reader, especially in dense scenes, which can better ensure the high quality and continuity of data collection.

[0067] S3: If the fluctuation range of the signal strength value is large, the signal propagation time value of each RFID tag is recorded, and the deviation of the signal propagation time value in the same time window is determined.

[0068] The signal propagation time value refers to the time interval from the RFID tag transmitting the signal to the card reader receiving the signal. If the signal strength value fluctuates greatly, the high-precision time synchronization mechanism of the card reader is used to record the signal time characteristic (TOF) value of each RFID tag, usually in microseconds or nanoseconds. The TOF values ​​of multiple tags in the same time window are measured one by one.

[0069] Analyze the TOF values ​​of multiple tags in the same time window and calculate the propagation time deviation of each tag , the calculation expression is: ; Among them, the TOF reference value is the reference time value set by the system (such as the average value or historical normal value).

[0070] The TOF propagation deviation index is generated by analyzing the deviation of the signal propagation time value in the same time window. The method for obtaining the TOF propagation deviation index is as follows:

[0071] A sequence of TOF values ​​collected in the same time window: ; Where M is the total number of RFID tags in the current time window, each TOF value represents the signal propagation time of a tag, and the TOF value sequence is sorted from small to large, recorded as ; Calculate the quartiles: the first quartile (Q1): the TOF value in the lower quartile (25% position) after sorting, the third quartile (Q3): the TOF value in the upper quartile (75% position) after sorting. Interquartile range (IQR): IQR=Q3−Q1; The upper and lower limits are set as: Lower limit: ; Upper bound: ; Mark outliers: TOF values ​​that are less than the lower bound and greater than the upper bound are marked as outliers, and the proportion of outliers to the total TOF values ​​in the same time window is calculated, that is, the TOF propagation deviation index is calculated.

[0072] When the TOF propagation deviation index is large, it means that there is a significant deviation or outlier in the signal propagation time value (TOF) within the same time window. This usually indicates that there is a serious conflict or interference between the signals of multiple tags, and the reader has difficulty in correctly distinguishing the signals of each tag, resulting in an increase in the decoding error rate. Therefore, the larger the TOF propagation deviation index, the lower the decoding accuracy of the reader in a high-density tag environment may be, and it is easy to misread, miss read or data overlap.

[0073] When the TOF propagation deviation index is small, it means that the TOF values ​​in the same time window are concentrated, the signal propagation time is relatively consistent, and the proportion of outliers is low. This indicates that there is less interference with the tag signal, and the reader can decode the information of each tag more accurately, thereby improving the decoding success rate and data reliability. Therefore, the smaller the TOF propagation deviation index, the higher the decoding accuracy of the reader, especially in complex scenarios, which can better ensure data quality.

[0074] S4: Evaluate the accuracy of data information decoded by the card reader for multiple tags based on the fluctuation amplitude of the signal strength value of the RFID tag and the deviation of the signal propagation time value within the same time window.

[0075] The TOF propagation deviation index and the RSSI density distribution fluctuation index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the accuracy value label of the data information decoded by the card reader for multiple tags as the prediction target, and takes minimizing the sum of the prediction errors of the accuracy value labels of the data information decoded by all readers for multiple tags as the training target. The machine learning model is trained until the sum of the prediction errors converges, and the model training is stopped. The accuracy value of the data information decoded by the card reader for multiple tags is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0076] The method for obtaining the accuracy value of the data information decoded by the card reader for multiple tags is as follows: from the comprehensive feature vector training data of the trained machine learning model, the corresponding function expression is obtained: ; In the formula, is the output function of the model, HK is the TOF propagation deviation index, is the RSSI density distribution fluctuation index, The accuracy value of the data information decoded by the reader for multiple tags.

[0077] S5: According to the evaluation result of the accuracy of the data information of the multiple tags decoded by the card reader, the decoding results of the card reader are divided into accurate decoding results, incompletely accurate decoding results and inaccurate decoding results, and corresponding management is performed.

[0078] Compare the acquired accuracy value of the data information of the multiple tags decoded by the card reader with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and compare the accuracy value of the data information of the multiple tags decoded by the card reader with the first standard threshold and the second standard threshold respectively;

[0079] If the accuracy value of the data information decoded by the card reader for multiple tags is greater than the second standard threshold, it means that the accuracy of the data information decoded by the card reader for multiple tags is high, and a high-accuracy decoding signal is generated at this time, and the decoding result of the card reader is classified as an accurate decoding result;

[0080] If the accuracy value of the data information decoded by the card reader for multiple tags is greater than or equal to the first standard threshold value and less than or equal to the second standard threshold value, it means that the accuracy of the data information decoded by the card reader for multiple tags is average, and a medium accuracy decoding signal is generated at this time, and the decoding result of the card reader is classified as an incomplete accuracy decoding result;

[0081] If the accuracy value of the data information decoded by the card reader for multiple tags is less than the first standard threshold, it means that the accuracy of the data information decoded by the card reader for multiple tags is low. At this time, a low-accuracy decoding signal is generated, and the decoding result of the card reader is classified as an inaccurate decoding result.

[0082] Store high-accuracy decoding results directly in the core database for subsequent data analysis and model optimization. Record relevant parameters (such as tag ID, TOF index, RSSI index, timestamp, etc.) and mark them as high-confidence data. High-accuracy decoding results are displayed first in the real-time monitoring system to provide users with reliable tag decoding information. Used to generate real-time motion trajectories, statistical athlete data or guide game decisions. Use high-accuracy data as training samples for machine learning models to further improve the prediction performance of the model.

[0083] For decoding results with incomplete accuracy, compare the high-accuracy decoding results of the previous and next time windows, and perform interpolation correction on the inaccurate data; use machine learning models or motion trajectory prediction models to re-estimate missing or deviated decoding values.

[0084] Store inaccurate decoding results in an independent error data pool, separate from the core database, to prevent them from affecting overall data analysis. Record the error type (such as signal loss, decoding failure, etc.) and possible causes. Analyze the fluctuation characteristics of RSSI and TOF data to locate the main factors that cause decoding failure (such as interference sources, tag failure). Check whether there are signal conflicts caused by too many tags or too high density. When the proportion of inaccurate decoding results exceeds the set threshold, send an alert to the system administrator, indicating that the reader configuration or signal processing algorithm may need to be adjusted. Provide optimization suggestions, such as adjusting the reader's read power, optimizing the anti-collision algorithm, or redeploying hardware. Use inaccurate decoding results for adversarial training to improve the machine learning model's ability to predict abnormal situations. Introduce anomaly detection mechanisms to identify potential problems in real time during future decoding processes.

[0085] S6: For decoding results with incomplete accuracy, the positions and times of the known points of the two RFID tags before and after are used for interpolation calculations, and the kinematic model is dynamically adjusted according to the calculation results and the athlete's movement pattern to optimize the accuracy of the data decoding results.

[0086] For the decoding result with incomplete accuracy, the position and time of the two known points before and after are used to calculate the position of the current time point. The calculation expression is: ; In the formula, is the predicted position corresponding to the current time point t (decoding result after interpolation), is the location of the previous known point of the RFID tag, After the RFID tag knows the location of the point, is the timestamp of the previous known point, is the timestamp of the next known point, t is the timestamp of the location point to be predicted, satisfying < < .

[0087] When the athlete's movement pattern changes (such as acceleration, deceleration or cornering), the interpolation calculation needs to be dynamically adjusted in combination with the kinematic model. The uniform motion model is suitable for situations where the athlete's speed is constant or the speed change in a short period of time is negligible. The expression is: ; where v is the speed of the athlete (calculated by the position and time of two known points before and after), ; The uniform acceleration motion model is applicable to the stage of athlete acceleration or deceleration, and the calculation expression is: ; In the formula, a is the athlete's acceleration, which is calculated based on the speed change between the two known points before and after. ; is the velocity at the previous moment, ; The curvilinear motion model is applicable to athletes moving along a curve or non-straight trajectory. The calculation expression is: ; In the formula, , , where is the coordinate of the center of the arc, r is the radius of the curve, which is determined by the geometric relationship between the two positions before and after. is the center angle of the athlete at the current time point, ω is the angular velocity of the athlete, is the predicted position at the current time point.

[0088] Combined with the motion mode, the appropriate model is selected according to the dynamic conditions for interpolation calculation. The comprehensive formula is expressed as: ;

[0089] The current position is directly calculated through the two known points before and after, and the uniform speed, acceleration and curve motion modes are judged according to the athlete's speed change and trajectory characteristics; based on the motion mode, a suitable kinematic model is selected, the interpolation formula is adjusted, the decoding result is optimized, and the optimized decoding result is compared with the adjacent known points to judge the optimization effect. If the optimization effect is good, it will be stored in the database.

[0090] Compare the optimized decoding results with the adjacent known points to determine the optimization effect, specifically:

[0091] Calculate the time deviation between the optimized decoding result and the adjacent known points. The expression is: ; In the formula, is the time point of the optimized decoding result, is the time point adjacent to the known point. If Δt exceeds the threshold set by the system, it is considered that the optimization effect may be poor.

[0092] Calculate the distance deviation between the optimized decoding result and the adjacent known points in space: ; is the position coordinate of the optimized decoding result, is the position coordinate of the adjacent known point. If Δd is less than the spatial error threshold set by the system, the optimization result is considered reliable.

[0093] The optimization effect can be judged by comparing the trajectory continuity of the optimized decoding results with the known points before and after: ; In the formula, To optimize the distance between the result and the known points before and after, is the time interval between the previous and next known points. Should be less than the time threshold Tthresh, Δd should be less than the spatial error threshold Dthresh, if the trajectory continuity C is within the preset standard range of the athlete's actual movement speed , the optimization effect is considered good.

[0094] The optimization effect is classified according to the conditions met: Good effect: All evaluation conditions are met and the optimization result is reliable. Average effect: Some conditions are met and the optimization result can be used as temporary data, but it needs to be marked for further verification. Poor effect: The conditions are not met and the optimization result is unavailable.

[0095] Good results: Mark the optimization results as highly reliable data and store them in the core database for subsequent analysis and modeling. Fair results: Store the optimization results in a temporary database and mark them as “data to be verified” for subsequent verification. Poor results: Isolate the optimization results to the error data pool for error analysis and algorithm optimization.

[0096] Optimization result data: including timestamp, position coordinates (x, y), optimized parameters (such as speed, acceleration). Comparison evaluation indicators: record the specific values ​​of Δt, Δd, C. Mark as "high confidence", "to be verified" or "wrong data". Use high confidence optimization results to retrain the model to improve subsequent prediction accuracy. Feed the data to be verified back to the model for cross-validation to improve the ability to handle complex situations. Display the decoding results with good optimization effects on the system monitoring interface, and adjust the system parameters (such as error threshold, model weight) in real time.

[0097] In this embodiment, by selecting several RFID tags with unique IDs, a test site for intensive multi-tag activities is constructed, and the number of tags and the range of movement speed are set. The number of tags is gradually increased and the distribution density is adjusted within the coverage area of ​​the card reader to simulate a high-density conflict environment, and the signal strength value is recorded and its fluctuation range is analyzed. If the signal strength fluctuation range is large, the signal propagation time value (TOF) is recorded, and the deviation of the TOF value in the same time window is determined. Based on the signal strength fluctuation range and TOF deviation, the accuracy of the card reader decoding multiple tag data information is evaluated. According to the evaluation results, the decoding results are divided into three categories: high accuracy, incomplete accuracy and low accuracy, and are classified and managed separately. For the incomplete accuracy decoding results, the position and time of the known points before and after are used for interpolation calculation, and the kinematic model is dynamically adjusted in combination with the athlete's movement pattern to optimize the accuracy of the decoding results.

[0098] Embodiment 2, a smart sports data analysis and management system based on the Internet of Things described in this embodiment includes a test environment construction module, a signal strength acquisition and analysis module, a signal propagation time analysis module, a decoding accuracy evaluation module, a decoding result classification management module and a decoding result optimization module:

[0099] Test environment construction module: select several RFID tags with unique IDs, build a test site with intensive multi-tag activity, and set the number of RFID tags and the range of movement speed in the test site;

[0100] Signal strength collection and analysis module: In the reader coverage area, gradually increase the number of tags and control the distribution density of tags to simulate the conflict environment of high-density scenes. Under different numbers and densities of tags, record the signal strength value of each tag and analyze the fluctuation range of the signal strength value.

[0101] Signal propagation time analysis module: If the fluctuation range of the signal strength value is large, the signal propagation time value of each RFID tag is recorded, and the deviation of the signal propagation time value in the same time window is determined;

[0102] Decoding accuracy evaluation module: evaluates the accuracy of data information decoded by the card reader for multiple tags based on the fluctuation of the signal strength value of the RFID tag and the deviation of the signal propagation time value in the same time window;

[0103] Decoding result classification management module: according to the evaluation results of the accuracy of the data information of multiple tags decoded by the card reader, the decoding results of the card reader are divided into accurate decoding results, incompletely accurate decoding results and inaccurate decoding results, and corresponding management is performed;

[0104] Decoding result optimization module: For decoding results with incomplete accuracy, the positions and times of the known points of the two front and rear RFID tags are used for interpolation calculations, and the kinematic model is dynamically adjusted according to the calculation results and the athlete's movement pattern to optimize the accuracy of the data decoding results.

[0105] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A smart sports data analysis and management method based on the Internet of Things, characterized by: The following steps are involved: S1: Select several RFID tags with unique IDs, build a test site with intensive multi-tag activities, and set the number of RFID tags and the range of movement speed in the test site; S2: In the reader coverage area, gradually increase the number of tags and control the distribution density of tags to simulate the conflict environment of high-density scenarios. Under different numbers and densities of tags, record the signal strength value of each tag and analyze the fluctuation range of the signal strength value. S3: If the fluctuation range of the signal strength value is large, the signal propagation time value of each RFID tag is recorded, and the deviation of the signal propagation time value in the same time window is determined; S4: Evaluate the accuracy of data information decoded by the card reader for multiple tags based on the fluctuation amplitude of the signal strength value of the RFID tag and the deviation of the signal propagation time value in the same time window; S5: according to the evaluation result of the accuracy of the data information of the multiple tags decoded by the card reader, the decoding results of the card reader are divided into accurate decoding results, incompletely accurate decoding results and inaccurate decoding results, and corresponding management is performed; S6: For decoding results with incomplete accuracy, the positions and times of the known points of the two RFID tags before and after are used for interpolation calculations, and the kinematic model is dynamically adjusted according to the calculation results and the athlete's movement pattern to optimize the accuracy of the data decoding results.

2. According to the method for analyzing and managing smart sports data based on the Internet of Things according to claim 1, it is characterized by: In S2, the signal strength value of each tag is recorded, and the fluctuation amplitude of RSSI data with different distribution densities is analyzed to generate the RSSI density distribution fluctuation index. The RSSI density distribution fluctuation index is obtained as follows: Collect RSSI signal strength value x(s) and record its sequence changing over time ; Sample the RSSI signal at a fixed sampling frequency to obtain equally spaced discrete time point data. N is the total length of the RSSI signal sequence. Select the window function w(s), set the window length L, and divide the signal strength value x(s) into multiple segments on the time axis according to the window function w(s). Each segment is L in length and weighted. For each segment, calculate its Fourier transform to obtain the frequency component of the signal. The expression is: ; where the window center is at τ, is the Fourier transform result of the signal strength value x(s), is the kernel function of Fourier transform, is the frequency. In discrete time signals, STFT is expressed as: ; Calculate the frequency components of each window Amplitude value , and record it, τ is the window center position on the time axis, is the discrete frequency component, representing the discrete points on the frequency axis, n is the number of windows, is the kernel function of the discrete transformation, recording all time windows τ and frequency components The amplitude , and obtain the time-frequency distribution matrix , calculate the total energy in each time window, the expression is: ; Represents the frequency energy density of the signal at time τ, counts the mean μE and standard deviation σE of energy fluctuations in all time windows, and calculates the RSSI density distribution fluctuation index, which is expressed as: ; In the formula, It is the RSSI density distribution fluctuation index.

3. According to the method for analyzing and managing smart sports data based on the Internet of Things according to claim 2, it is characterized in that: In S3, the deviation of the signal propagation time value in the same time window is analyzed to generate a TOF propagation deviation index. The TOF propagation deviation index is obtained as follows: Collect a sequence of TOF values ​​within the same time window ; Where M is the total number of RFID tags in the current time window, each TOF value represents the signal propagation time of a tag, and the TOF value sequence is sorted from small to large, recorded as ; Calculate the quartiles: the first quartile Q1 is the TOF value in the lower quartile after sorting, the third quartile Q3 is the TOF value in the upper quartile after sorting, and the interquartile range IQR: IQR=Q3−Q1; the upper and lower limits are set as: Lower limit: ; Upper bound: ; The TOF values ​​that are less than the lower bound and greater than the upper bound are recorded as outliers, and the proportion of outliers to the total TOF values ​​in the same time window is calculated, that is, the TOF propagation deviation index is calculated.

4. The method for analyzing and managing smart sports data based on the Internet of Things according to claim 3 is characterized in that: In S4, the accuracy of the data information decoded by the card reader for multiple tags is evaluated according to the fluctuation amplitude of the signal strength value of the RFID tag and the deviation of the signal propagation time value in the same time window, specifically: The TOF propagation deviation index and the RSSI density distribution fluctuation index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the accuracy value label of the data information decoded by the card reader for multiple tags as the prediction target, and takes minimizing the sum of the prediction errors of the accuracy value labels of the data information decoded by all readers for multiple tags as the training target. The machine learning model is trained until the sum of the prediction errors converges, and the model training is stopped. The accuracy value of the data information decoded by the card reader for multiple tags is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

5. The method for analyzing and managing smart sports data based on the Internet of Things according to claim 4 is characterized in that: In S5, according to the evaluation result of the accuracy of the data information of the multiple tags decoded by the card reader, the decoding result of the card reader is divided into an accurate decoding result, an incompletely accurate decoding result and an inaccurate decoding result, specifically: Compare the acquired accuracy value of the data information of the multiple tags decoded by the card reader with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and compare the accuracy value of the data information of the multiple tags decoded by the card reader with the first standard threshold and the second standard threshold respectively; If the accuracy value of the data information decoded by the card reader for multiple tags is greater than the second standard threshold, it means that the accuracy of the data information decoded by the card reader for multiple tags is high, and a high-accuracy decoding signal is generated at this time, and the decoding result of the card reader is classified as an accurate decoding result; If the accuracy value of the data information decoded by the card reader for multiple tags is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the accuracy of the data information decoded by the card reader for multiple tags is average, and a medium accuracy decoding signal is generated at this time, and the decoding result of the card reader is classified as an incomplete accuracy decoding result; If the accuracy value of the data information decoded by the card reader for multiple tags is less than the first standard threshold, it means that the accuracy of the data information decoded by the card reader for multiple tags is low. At this time, a low-accuracy decoding signal is generated, and the decoding result of the card reader is classified as an inaccurate decoding result.

6. The method for analyzing and managing smart sports data based on the Internet of Things according to claim 5 is characterized in that: In S6, for the incomplete accuracy decoding result, the positions and times of the known points of the two RFID tags before and after are used for interpolation calculation, and the kinematic model is dynamically adjusted according to the calculation results and the athlete's movement pattern, specifically: For the decoding result with incomplete accuracy, the position and time of the two known points before and after are used to calculate the position of the current time point. The calculation expression is: ; In the formula, is the predicted position corresponding to the current time point t, is the location of the previous known point of the RFID tag, After the RFID tag knows the location of the point, is the timestamp of the previous known point, is the timestamp of the next known point, t is the timestamp of the location point to be predicted, satisfying < < ; When the athlete's movement pattern changes, the interpolation calculation needs to be dynamically adjusted in combination with the kinematic model. The uniform motion model expression is: ; where v is the speed of the athlete, ; The calculation expression of uniform acceleration motion model is: ; In the formula, a is the athlete's acceleration, which is calculated based on the speed change between the two known points before and after. ; is the velocity at the previous moment, ; The calculation expression of the curve motion model is: ; In the formula, , , where is the coordinate of the center of the arc, r is the radius of the curve, which is determined by the geometric relationship between the two positions before and after. is the center angle of the athlete at the current time point, ω is the angular velocity of the athlete, is the predicted position at the current time point; Combined with the motion mode, interpolation calculation is performed according to dynamic conditions, and the comprehensive formula is expressed as: ; The current position is directly calculated through the two known points before and after, and the uniform speed, acceleration and curve movement modes are judged according to the athlete's speed change and trajectory characteristics; the kinematic model is selected based on the movement mode, the interpolation formula is adjusted, the decoding result is optimized, and the optimized decoding result is compared with the adjacent known points to judge the optimization effect. If the optimization effect is good, it will be stored in the database.

7. The method for analyzing and managing smart sports data based on the Internet of Things according to claim 6, characterized in that: Compare the optimized decoding results with the adjacent known points to determine the optimization effect, specifically: Calculate the time deviation between the optimized decoding result and the adjacent known points. The expression is: ; In the formula, is the time point of the optimized decoding result, is a time point adjacent to a known point; Calculate the distance deviation between the optimized decoding result and the adjacent known points in space: ; is the position coordinate of the optimized decoding result, is the position coordinates of the nearby known points; The optimization effect can be judged by comparing the trajectory continuity of the optimized decoding results with the known points before and after: ; In the formula, To optimize the distance between the result and the known points before and after, is the time interval between the previous and next known points, Should be less than the time threshold Tthresh, Δd should be less than the spatial error threshold Dthresh, if the trajectory continuity C is within the preset standard range of the athlete's actual movement speed , the optimization effect is considered good.

8. A smart sports data analysis and management system based on the Internet of Things, used to implement a smart sports data analysis and management method based on the Internet of Things as described in any one of claims 1 to 7, characterized in that: It includes test environment construction module, signal strength acquisition and analysis module, signal propagation time analysis module, decoding accuracy evaluation module, decoding result classification management module and decoding result optimization module: Test environment construction module: select several RFID tags with unique IDs, build a test site with intensive multi-tag activity, and set the number of RFID tags and the range of movement speed in the test site; Signal strength collection and analysis module: In the reader coverage area, gradually increase the number of tags and control the distribution density of tags to simulate the conflict environment of high-density scenes. Under different numbers and densities of tags, record the signal strength value of each tag and analyze the fluctuation range of the signal strength value. Signal propagation time analysis module: If the fluctuation range of the signal strength value is large, the signal propagation time value of each RFID tag is recorded, and the deviation of the signal propagation time value in the same time window is determined; Decoding accuracy evaluation module: evaluates the accuracy of data information decoded by the card reader for multiple tags based on the fluctuation of the signal strength value of the RFID tag and the deviation of the signal propagation time value in the same time window; Decoding result classification management module: according to the evaluation results of the accuracy of the data information of multiple tags decoded by the card reader, the decoding results of the card reader are divided into accurate decoding results, incompletely accurate decoding results and inaccurate decoding results, and corresponding management is performed; Decoding result optimization module: For decoding results with incomplete accuracy, the positions and times of the known points of the two front and rear RFID tags are used for interpolation calculations, and the kinematic model is dynamically adjusted according to the calculation results and the athlete's movement pattern to optimize the accuracy of the data decoding results.

Images