Multi-modal wearable system for force synergy quantitative evaluation of martial arts training and competition, computer device and storage medium

Abstract

The present application relates to the technical fields of sports biomechanics, wearable sensing and biosignal processing, and discloses a multi-modal wearable system for force coordination quantitative evaluation of martial arts training and competition, a computer device and a storage medium. The system synchronously collects the pulse signal and IMU sensing data of athletes through a sensing module. The sequence extraction unit of the force coordination quantitative evaluation module obtains the heart rate variability sequence and the action force sequence respectively. Then, the force coordination degree, the force concentration degree and the action stability are calculated by the Pearson correlation coefficient, the ratio of the acceleration peak value and the standard deviation, and the Euclidean distance error respectively. Finally, the three scores are weighted and fused by the comprehensive evaluation unit, and the total evaluation score is output. The present application realizes the objective quantitative evaluation of force coordination, force concentration and action stability of martial arts, and provides reliable data support for professional evaluation of martial arts events.

Description

Technical Field

[0001] This invention relates to the interdisciplinary fields of wearable sensing, biosignal processing, and quantitative evaluation of martial arts competitions. In particular, it relates to a multimodal wearable system, computer equipment, and storage medium for quantitative evaluation of qi-strength coordination in martial arts training and competition. Background Technology

[0002] The current application of biosensor technology in the field of martial arts suffers from the following technical shortcomings: 1. The sensing dimension is too singular to match the core evaluation logic of the unity of form and energy in martial arts.

[0003] Existing technologies are mostly single-sensor modes, only collecting striking data or movement shape data. Common solutions only collect basic physiological data and lack a multimodal sensing architecture that combines physiological and biomechanical aspects. They cannot simultaneously acquire data related to energy coordination and breathing, and can only assess the shape of movements, failing to reflect differences in skill level. Furthermore, they have not established a mechanism for the coordinated acquisition of physiological and biomechanical indicators.

[0004] 2. The sensor signal has poor adaptability and cannot meet the needs of collecting explosive movements in martial arts.

[0005] Existing wearable sensing devices and general filtering algorithms cannot suppress motion artifacts caused by instantaneous force exertion in martial arts, which easily leads to PPG signal distortion. Secondly, the sampling rate is fixed and cannot take into account both power accumulation and power exertion scenarios, resulting in low data signal-to-noise ratio. Furthermore, its signal processing and sampling strategies are only suitable for periodic aerobic exercise and are not suitable for the non-periodic movement characteristics of martial arts power accumulation, explosive power, and recovery.

[0006] 3. Insufficient timing synchronization accuracy of multi-source sensors makes pneumatic coordination analysis prone to distortion.

[0007] Existing solutions do not involve martial arts-specific timing synchronization mechanisms. The sampling rates of physiological and movement data are mismatched, and simple alignment based solely on general timestamps cannot achieve millisecond-level synchronization accuracy. The timing of physiological and movement data is misaligned at the moment of force exertion, resulting in distorted quantitative evaluation results of qi-force coordination, which cannot meet the high-precision alignment requirements of martial arts for short-term instantaneous force exertion.

[0008] 4. The lack of a quantitative model specific to martial arts prevents the practical application of the professional value of sensor data.

[0009] Existing equipment only collects basic data and lacks a quantitative analysis model specific to martial arts. Physiological and movement data can only be statistically analyzed and used for early warning, and cannot be converted into quantitative evaluation indicators such as energy coordination, power concentration, and movement stability. This is disconnected from the professional martial arts evaluation system and makes it difficult to support professional competition evaluation.

[0010] 5. The lack of a competition-level secure access and transmission mechanism poses a risk of data leakage.

[0011] Existing martial arts sensors mostly use general Bluetooth transmission, without hardware authentication and encrypted transmission. Device access is not whitelisted and data is not encrypted. In formal training and competition scenarios, there is a high risk of security problems such as leakage of athletes' sensitive physiological data and tampering with training data. This cannot meet the dual requirements of professional training systems for data security and competition fairness.

[0012] In summary, existing technologies suffer from several technical problems, including limited sensing dimensions, poor adaptability of sensing signals, insufficient timing synchronization accuracy of multi-source sensing, lack of a quantitative model specific to martial arts, lack of secure data access and transmission mechanisms, and high risk of data leakage. Summary of the Invention

[0013] To address the shortcomings of the existing technologies, this invention provides a multimodal wearable system, computer equipment, and storage medium for quantitative evaluation of qi and strength coordination in martial arts training and competition. The technical solution of the system provided by this invention includes: Sensing module: used to simultaneously acquire the athlete's pulse signal and IMU sensor data; Collaborative Quantitative Evaluation Module: Connected to the sensing module, including: Sequence extraction unit: used to extract heart rate variability sequence and force exertion sequence based on the pulse signal and the IMU sensor data, respectively; Strength coordination acquisition unit: connected to the sequence extraction unit, used to align the heart rate variability sequence and the force exertion sequence, calculate the Pearson correlation coefficient, and obtain the strength coordination score; Force concentration acquisition unit: connected to the sequence extraction unit, used to calculate the ratio of the peak acceleration of each action to the standard deviation of the acceleration sequence based on the force intensity sequence of the action, and obtain the force concentration score; Action stability acquisition unit: connected to the sequence extraction unit, used to calculate the Euclidean distance error between the action force sequence and the preset action sequence, and obtain the action stability score; The comprehensive evaluation unit is connected to the force coordination acquisition unit, the force concentration acquisition unit, and the motion stability acquisition unit, respectively, and is used to weight and fuse the force coordination score, the force concentration score, and the motion stability score to obtain the total evaluation score.

[0014] Preferably, the system further includes a preprocessing module connected to the sensing module, wherein the preprocessing module performs preprocessing in the following ways: S1. Filter the pulse signal and the IMU sensing data respectively; and use the acceleration signal in the IMU sensing data as reference noise to eliminate motion interference in the pulse signal through an adaptive noise cancellation algorithm; to obtain the filtered pulse signal and IMU sensing data; S2. Outlier identification and removal are performed on the filtered pulse signal and the IMU sensing data; signal completion is performed using linear interpolation and spline fitting methods to obtain the completed pulse signal and IMU sensing data; S3. Extract the heart rate variability sequence and the force exertion sequence based on the completed pulse signal and the IMU sensor data, respectively; and perform time synchronization of the heart rate variability sequence and the force exertion sequence through a preset time synchronization mechanism.

[0015] Preferably, the preset timing synchronization mechanism includes: S301. Based on the NTP network time protocol, complete the clock reference calibration between the sensing module and the gateway, and output a unified clock source; S302. Based on the unified clock source, add timestamps to the heart rate variability sequence and the force exertion sequence, and output the heart rate variability sequence and the force exertion sequence with timestamps; S303. The dynamic time warping algorithm is used to align the timestamped heart rate variability sequence and the force exertion sequence.

[0016] Preferably, the system further includes a sensor access and transmission module connected to the sensing module. The implementation method of the sensor access and transmission module includes: pre-setting a unique hardware ID and key for each sensing unit in the sensing module; the gateway has a built-in hardware encryption unit and a device whitelist; when the sensing unit is accessed, the gateway verifies the hardware ID and compares the key, and then assigns a temporary session ID. The temporary session ID will expire after a single event ends. The data collected by the sensing module is encrypted using the national cryptographic SM4 algorithm and transmitted to the gateway based on the MQTT lightweight publish-subscribe protocol. Physiological data, action data, and device status data are classified and transmitted separately through independent topics. The data packetization strategy and transmission frequency of the sensing module are optimized, and data frame compression technology is used to compress the length of a single packet of data.

[0017] Preferably, the sensing module includes: Physiological sensing unit: used to collect data including the athlete's heart rate, blood oxygen saturation, pulse wave velocity, and heart rate variability. Motion biomechanics sensing unit: used to collect data including the athlete's motion acceleration, angular velocity, and posture angle.

[0018] Preferably, the system further includes a data visualization and basic management module: connected to the collaborative quantitative evaluation module, used to provide a web-based visualization interface, including: real-time display of athletes' physiological indicator curves, force curves, power radar charts, and total evaluation scores; used to support local storage, retrieval, and playback of competition data, as well as export of analysis reports in PDF / Excel format.

[0019] The present invention also provides a computer device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to realize the function of the multimodal wearable system for quantitative evaluation of qi and strength coordination for martial arts training and competition as described in any of the above-described contents.

[0020] The present invention also provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores computer instructions, the computer instructions being used to enable a computer to implement the functions of the multimodal wearable system for quantitative evaluation of qi and strength coordination for martial arts training and competition as described in any of the above-described contents.

[0021] Beneficial effects The technical solution of this invention fills the gap in the quantitative sensing technology of martial arts skills, realizing dual data support for the "form" and "quality" of martial arts movements; it significantly improves the signal-to-noise ratio and reliability of sensing data; it achieves millisecond-level signal alignment based on a three-level time synchronization mechanism, adapting to the instantaneous force exertion characteristics of martial arts and ensuring the accuracy of quantitative analysis; it constructs a martial arts-specific energy coordination quantitative model, transforming traditional subjective skill evaluation into objective quantitative scoring; it is equipped with hardware identity authentication and a national cryptographic encryption transmission system, balancing data security and low-latency transmission performance; at the same time, it adopts a lightweight wearable terminal, which is convenient and flexible to wear, quick to deploy, and highly adaptable to different sites, greatly reducing equipment investment costs. Attached Figure Description

[0022] 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.

[0023] Figure 1 This is a schematic diagram of a multimodal wearable system for quantitative evaluation of qi and strength coordination in martial arts training and competition, provided in a preferred embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of a computer device provided by the present invention. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Example 1 like Figure 1 As shown, the present invention provides a multimodal wearable system for quantitative evaluation of qi and strength coordination in martial arts training and competition.

[0026] The existing technologies in the field of intelligent assessment of martial arts movements can be mainly divided into the following three categories: For example, the invention patent (publication number: CN117695597A) relies on pressure-sensing protective gear and ring sensors to collect and score data on striking and contact in competitions; the invention patent (publication number: CN118470750A) is based on visual skeletal key point detection and only completes the appearance similarity score through movement trajectory and posture comparison; the invention patent (publication number: CN117653049A) is a conventional wearable biosensor system, which is only suitable for health monitoring of physiological and biochemical signals in steady-state scenarios and is not adapted to martial arts sports scenarios.

[0027] Existing technologies generally suffer from numerous shortcomings: They are limited to a single sensing dimension, confined to impact data or visual shape acquisition, lacking a multimodal synchronous acquisition architecture encompassing physiological and biomechanical aspects, thus failing to align with the evaluation logic of martial arts' unity of form and energy, and coordinated force; they haven't optimized signal processing for the explosive, non-periodic force characteristics of martial arts, making it difficult to suppress motion artifacts, resulting in distorted physiological signals and poor acquisition adaptability; they lack a dedicated temporal synchronization mechanism, leading to insufficient alignment accuracy of multi-source data from different frequencies, and temporal misalignment of physiological and movement characteristics at the moment of force application, resulting in distorted results in the coordinated force analysis; they lack a martial arts-specific quantitative model, only enabling basic data acquisition and simple statistics, unable to transform into professionally scoreable quantitative indicators of skill; and they lack encrypted transmission and access control mechanisms for data, posing security risks such as leakage of sensitive physiological data of athletes and data tampering, failing to meet the data security and reliable management requirements of professional competitions or training systems.

[0028] To address the problems existing in the prior art, in some preferred embodiments, the system of the present invention includes a sensing module and a collaborative quantitative evaluation module.

[0029] The sensing module is used to simultaneously acquire the athlete's pulse signal and IMU sensor data. In some preferred embodiments, the sensing module of the present invention includes a physiological sensing unit and a movement biomechanical sensing unit. The physiological sensing unit uses a wearable smart ring with an integrated multi-wavelength PPG photoplethysmography sensor as the core acquisition terminal. It has a built-in green / infrared dual emitter, a high-sensitivity photodiode, a low-noise signal conditioning circuit, and a force detection subunit. It can acquire the athlete's core physiological indicators such as heart rate, blood oxygen saturation, pulse wave velocity, and heart rate variability. The basic sampling rate is 50Hz, which can be automatically increased to 100Hz during the force exertion phase of the movement. The device weighs ≤8g, supports continuous wear throughout the entire movement, and will not affect the athlete's performance. The motion biomechanics sensing unit employs a distributed six-axis IMU (integrated accelerometer and gyroscope) to target the core movement characteristics of martial arts punches, kicks, body movements, and footwork. Sensing points are deployed at six key joints of the athlete: wrist, elbow, shoulder, hip, knee, and ankle. Each node weighs ≤5g and is secured with elastic straps, allowing no restriction on joint movement. The sensors collect core biomechanical data such as the athlete's acceleration, angular velocity, and posture angles, with a base sampling rate of 500Hz, which automatically increases to 1000Hz at the moment of force exertion. The force detection subunit monitors the amplitude and rate of change of the IMU acceleration signal in real time. When the peak acceleration or rate of change exceeds a preset force threshold, a sampling rate switching command is automatically triggered, increasing the sampling rate of the physiological sensing unit to 100Hz and the motion biomechanics sensing unit to 1000Hz. After the force exertion phase ends, the sampling rate of each sensing unit automatically returns to the base sampling rate, achieving adaptive dynamic adjustment of the sampling rate to meet the high-precision acquisition requirements of instantaneous force exertion in martial arts.

[0030] In some preferred embodiments, the sensors in the sensing module of the present invention further include: all sensors have a built-in high-precision RTC real-time clock module, a low-power MCU main control unit, and a wireless communication module, supporting dynamic adaptive adjustment of the sampling rate; at the same time, an expansion interface for an electromyography (EMG) sensor is reserved to be compatible with the acquisition of more dimensions of biological signals. The signals acquired by the sensing module are used for subsequent processing.

[0031] Collaborative quantitative evaluation module: connected to the sensing module, including: a pneumatic coordination degree acquisition unit, a force concentration degree acquisition unit, a motion stability acquisition unit, and a comprehensive evaluation unit.

[0032] Sequence extraction unit: used to extract heart rate variability sequence and force exertion sequence based on the pulse signal and the IMU sensor data, respectively. In some preferred embodiments, the sequence extraction unit of the present invention receives the pulse signal and IMU sensor data output by the sensing module, analyzes the pulse signal to extract the heart rate variability sequence; simultaneously, it performs calculation and analysis on the IMU sensor data to extract and generate the force exertion sequence for subsequent processing.

[0033] The energy coordination acquisition unit, connected to the sequence extraction unit, aligns the heart rate variability sequence and the movement force intensity sequence, calculates the Pearson correlation coefficient, and obtains an energy coordination score. The heart rate variability sequence is a time-series data sequence characterizing the fluctuations in an athlete's physiological heart rhythm; the movement force intensity sequence is a mechanical time-series data sequence characterizing the instantaneous force intensity of a martial arts move over time. The core concept of "unity of energy and strength" in traditional martial arts is "using energy to control force, and force following energy," requiring athletes to maintain a high degree of synchronization between their breathing rhythm, physiological state, and the mechanical characteristics of the movement at the moment of force exertion. High-level athletes can regulate their autonomic nervous system through breathing, causing regular fluctuations in heart rate variability at the moment of force exertion, achieving precise release of physiological energy. Amateur athletes, however, often suffer from a "disconnect between energy and strength," where physiological changes are not synchronized with the force exertion, resulting in low force efficiency.

[0034] Existing technologies can only collect single-dimensional striking force data, without addressing the linkage analysis between physiological and biomechanical indicators. To address these technical issues, in some preferred embodiments, the heart rate variability sequence described in this invention can accurately reflect the regulatory state of the athlete's autonomic nervous system, indirectly characterizing changes in respiratory rhythm and physiological energy reserves; the movement force intensity sequence extracted by the IMU can accurately reflect the biomechanical output characteristics of the movement. In some preferred embodiments, this invention uses a dynamic time warping algorithm to match subsequences and temporal features between the two time series, ensuring a precise correspondence between physiological rhythm fluctuation periods and limb force exertion periods; then, the Pearson correlation coefficient algorithm is used to calculate the linear correlation between the trend of heart rate variability and the trend of movement force intensity; after normalizing the calculated Pearson correlation coefficient, a 0-100 point score for breath coordination is output. The Pearson correlation coefficient between two sequences can quantify the degree of synchronization between changes in physiological state and mechanical output over time: the higher the correlation coefficient, the better the coordination between physiological regulation and force exertion, and the higher the degree of "unity of qi and strength"; the lower the correlation coefficient, the more likely there is a problem of "disconnection between qi and strength".

[0035] Force Concentration Acquisition Unit: Connected to the sequence extraction unit, this unit calculates the ratio of the peak acceleration to the standard deviation of the acceleration sequence for each movement based on the force exertion sequence, thus obtaining a force concentration score. Force concentration characterizes the efficiency of an athlete's martial arts movements; the acceleration sequence is the time-series data of the force exertion sequence; the peak acceleration refers to the maximum instantaneous amplitude of the three-axis combined acceleration output by the distributed six-axis IMU within a single martial arts movement cycle, directly reflecting the maximum explosive force of the instantaneous exertion of a martial arts move.

[0036] Traditional martial arts emphasize "whole-body power" and "explosive force," requiring the concentration of the body's strength into a single, instantaneous burst to avoid power dispersion. High-level athletes' power generation process is characterized by "short duration, high peak value, and low fluctuation," enabling them to transfer power to the target in an extremely short time. In contrast, amateur athletes' power generation process often lasts longer, with significant power fluctuations, exhibiting a "sluggish" and inefficient power generation. Current technology, which only collects the peak force at the moment of impact using pressure sensors, cannot reflect the overall characteristics of the power generation process and cannot effectively distinguish between "brute force" and "skillful force."

[0037] To address the aforementioned technical issues, in some preferred embodiments, this invention calculates the ratio of the peak acceleration to the standard deviation of the acceleration sequence for each movement, and outputs a force concentration score of 0-100 after normalization. The peak acceleration reflects the maximum force intensity of a single movement, while the standard deviation of the acceleration sequence reflects the degree of force fluctuation during the force exertion process. The ratio quantifies the degree of force concentration; a higher ratio indicates that the force is more concentrated at the peak moment, resulting in higher force efficiency and a better "whole-body power" effect; a lower ratio indicates that the force is severely dispersed during the force exertion process, leading to lower force efficiency. By using the ratio of peak to standard deviation, this invention simultaneously considers force intensity and the stability of the force exertion process, enabling a more accurate assessment of an athlete's force exertion technique level and effectively distinguishing the technical gaps between athletes of different levels.

[0038] Action stability acquisition unit: connected to the sequence extraction unit, used to calculate the Euclidean distance error between the action force sequence and the preset action sequence, and obtain the action stability score.

[0039] In traditional martial arts, the evaluation of "form" requires not only accurate posture but also that the rhythm, timing of force application, and transitions of movements conform to standardized norms. While the speed of the same movement may vary among different athletes, the temporal structure and posture changes must remain consistent. Current technology, employing visual skeletal keypoint detection, can only collect spatial positional information and is susceptible to environmental factors such as lighting and occlusion, resulting in low temporal accuracy.

[0040] To address the above issues, in some preferred embodiments, this invention aligns the force sequence of an action with a preset action sequence using a dynamic time warping algorithm. Alternatively, other reasonable algorithms can be designed by those skilled in the art based on actual conditions and site requirements. Then, the Euclidean distance error between the force sequence and the preset action sequence is calculated, normalized, and output as a 0-100 point action stability score. The action time sequence acquired by the IMU contains multi-dimensional mechanical information such as acceleration, angular velocity, and posture angles, comprehensively reflecting the dynamic characteristics of the action. The dynamic time warping algorithm solves the problem of inconsistent sequence lengths caused by differences in the speed of different athletes, achieving optimal alignment between the two time sequences. The Euclidean distance error of the aligned sequence quantifies the posture differences between the current action and the standard action at each time point, comprehensively reflecting the overall fit of the action: the smaller the error, the higher the action stability and the better the fit with the standard action. This invention uses IMU inertial measurement technology, which can acquire multi-dimensional mechanical time sequence information of the action with a timing accuracy of up to 1ms. Combined with the dynamic time warping algorithm, it can more accurately evaluate the dynamic characteristics and rhythmic consistency of the action, and the evaluation results are more in line with the requirements of professional martial arts judging.

[0041] A comprehensive evaluation unit, connected to the energy coordination acquisition unit, the force concentration acquisition unit, and the movement stability acquisition unit, is used to weight and fuse the energy coordination score, the force concentration score, and the movement stability score to obtain a total evaluation score. To address the lack of a dedicated quantitative model for martial arts in existing technologies, in some preferred embodiments, this invention deeply integrates biosensor data with the martial arts evaluation system to achieve quantitative evaluation of martial arts. The total evaluation score F is obtained through the following quantitative model: ;in, This represents the score for physical coordination. The score represents the concentration of effort exerted. This represents the score for movement stability. , , These are the weights of each item, and + + In some preferred embodiments, the weights of the present invention... , , The weights were determined by an expert group consisting of 5 national-level martial arts referees, 3 professors of sports training, and 2 biomechanics experts. The weights were calculated using the Delphi method after multiple rounds of independent scoring, followed by a consistency test (Cronbach's α > 0.85), and then by the analytic hierarchy process (AHP).

[0042] The "physiological + movement biomechanics" dual-modal biosensor architecture of this embodiment is specially designed for the professional martial arts evaluation system. It simultaneously collects athletes' internal physiological changes and external movement biomechanics data, breaking through the limitation of existing technologies that can only collect single-dimensional data. For the first time, it provides a feasible sensing and data collection solution for the core element of martial arts "qi-power coordination", realizing dual data support for the "form" (including external manifestations such as movement trajectory, posture, and rhythm) and "quality" (including internal skill quality such as qi-power coordination, force concentration, and movement stability) of martial arts movements.

[0043] Example 2 Based on Example 1, this example is used to preprocess the raw signals collected by the sensing module.

[0044] To address issues such as high-frequency noise in raw pulse signals and IMU sensor data, motion artifacts from explosive martial arts movements, signal abrupt changes, data gaps, and temporal misalignment in multi-source sensor data, this invention preprocesses the acquired raw signals. This avoids affecting the accuracy of subsequent pneumatic coordination feature extraction and quantitative evaluation due to raw signal distortion, data loss, and time asynchrony. In some preferred embodiments, the preprocessing module of this invention implements the following preprocessing method: S1. Filter the pulse signal and the IMU sensing data respectively; and use the acceleration signal in the IMU sensing data as reference noise to eliminate motion interference of the pulse signal through an adaptive noise cancellation algorithm; to obtain the filtered pulse signal and the IMU sensing data.

[0045] In some preferred embodiments, the present invention employs a fourth-order Butterworth filtering method to filter the pulse signal and IMU sensing data respectively, filtering out fundamental interference components such as power frequency interference and high-frequency random noise. Alternatively, other reasonable designs by those skilled in the art can be made based on actual conditions and site requirements. Then, the motion component of the IMU's built-in triaxial acceleration output is used as reference noise, and an adaptive noise cancellation algorithm is introduced to adaptively fit the artifact interference components caused by martial arts limb swaying and force exertion. The motion interference components are then canceled and stripped from the pulse signal, ultimately obtaining a pure pulse signal after noise reduction, purification, and removal of motion artifacts, and normalized IMU sensing data.

[0046] S2. Outlier identification and removal are performed on the filtered pulse signal and IMU sensing data; signal completion is performed using linear interpolation and spline fitting methods to obtain the completed pulse signal and IMU sensing data. In some preferred embodiments, the present invention performs full-domain traversal detection on the filtered pulse signal and IMU sensing data, uses the 3σ criterion to identify abrupt outliers and abnormal sampling points caused by sensor contact jitter and instantaneous environmental disturbances, and removes them; for the data missing segments formed after removing outliers, linear interpolation is used to smooth the short-interval missing segments, and spline fitting is used to curve-fit and restore the long-interval signal defects, repairing signal breakpoints, thereby ensuring that the entire time-series signal is continuous, smooth, and without discontinuities.

[0047] S3. Extract the heart rate variability sequence and the force exertion sequence based on the completed pulse signal and the IMU sensor data, respectively; align the heart rate variability sequence and the force exertion sequence in time using a preset timing synchronization mechanism. In some preferred embodiments, the present invention, based on a preset unified timing synchronization mechanism and using a standard clock reference, performs timestamp calibration and sequence registration on the filtered and completed pulse physiological data and IMU motion sensor data, correcting the timing offset caused by sampling frequency and acquisition start time difference between the two types of signals, achieving precise alignment of physiological signals and motion mechanical signals in the time dimension, ensuring a one-to-one correspondence between physiological changes and limb force exertion moments during subsequent pneumatic-force coordination feature analysis.

[0048] This embodiment can effectively filter out signal noise and motion interference caused by martial arts movements, significantly improve the signal-to-noise ratio and fidelity of pulse signals and IMU sensing data; remove abnormal data and repair signal loss, ensuring that the sensing time series data is complete and continuous and the change characteristics are real and effective; achieve precise time series matching between physiological signals and motion sensing data, eliminate time series offset and misalignment problems, and provide reliable basic data support for subsequent power quantitative assessment.

[0049] Example 3 Based on Embodiment 2, this embodiment provides a better preset timing synchronization mechanism.

[0050] Existing multi-source sensor data relies solely on local timestamps for simple alignment, resulting in issues such as clock asynchrony between sensor nodes and the gateway, low data time stamp accuracy, and large discrepancies in matching inter-frequency time sequences. This makes it difficult to meet the millisecond-level precision alignment requirements of instantaneous force exertion in martial arts scenarios, and easily leads to distortion in subsequent quantitative analysis of force coordination. To address these problems, in some preferred embodiments, the preset time synchronization mechanism provided by this invention includes: S301. Based on the NTP network time protocol, complete the clock reference calibration of the sensing module and the gateway, and output a unified clock source.

[0051] By relying on the NTP network time protocol to build a unified standard time reference for the entire domain, clock reference calibration is carried out on all distributed sensors and data gateways to correct the inherent drift and timing deviation of the local clock of each sensor, control the clock synchronization error to ≤100μs, and make the timing reference of all sensor acquisition nodes and gateways completely unified, thus avoiding timing deviations caused by differences in device clocks from the source.

[0052] S302. Based on the unified clock source, add timestamps to the heart rate variability sequence and the force exertion sequence, and output the heart rate variability sequence and the force exertion sequence with timestamps.

[0053] While collecting pulse signals and IMU sensing data, the sensing module binds a standardized timestamp that has been uniformly calibrated, and performs hardware-level real-time labeling on each frame of raw sensing data. The accurate timestamp serves as the basis for data time-series traceability, avoiding delays and deviations caused by subsequent software labeling, and ensuring that the timestamp of each piece of collected data is authentic, accurate and traceable.

[0054] S303. A dynamic time warping algorithm is used to align the timestamped heart rate variability sequence and the movement force intensity sequence. Addressing the differences in sampling frequencies and movement start and end times between the heart rate variability sequence and the movement force intensity sequence, a dynamic time warping algorithm is introduced to perform nonlinear stretching and matching calibration on the two time series sequences. This adaptively matches the temporal characteristics of physiological rhythm changes and limb force changes, eliminating time misalignment between different frequency and asynchronous sequences, and achieving temporal alignment of the two types of sequences.

[0055] This embodiment controls the synchronization error within 100μs and establishes a high-precision unified time reference; it avoids software marking delay deviation and ensures accurate and traceable data timing identification; it achieves millisecond-level correspondence between physiological changes and limb force exertion, with timing synchronization error controlled within 5ms, effectively solving the problem of timing misalignment of multi-source sensor data, and can provide high-precision timing data support for subsequent functional quantitative assessments.

[0056] Example 4 Based on Example 1, this example is used to achieve secure data access and transmission.

[0057] Existing martial arts competition sensing systems suffer from several problems, including lack of identity verification for terminal access, susceptibility to unauthorized device access, vulnerability to theft and tampering during plaintext data transmission, chaotic transmission of mixed data types, high data packet redundancy, high bandwidth consumption, and poor real-time performance. To address these issues, in some preferred embodiments, the system of this invention further includes a sensor access and transmission module connected to the sensing module. The implementation method of the sensor access and transmission module includes: pre-setting a unique hardware ID and key for each sensing unit in the sensing module; and a gateway with a built-in hardware encryption unit and device whitelist. When a sensing unit accesses the system, the gateway verifies the hardware ID and compares the key, then assigns a temporary session ID. This temporary session ID expires after each competition, thereby preventing unauthorized device access and ensuring secure and controllable terminal access.

[0058] The data collected by the sensing module is encrypted using the national cryptographic SM4 algorithm and transmitted to the gateway based on the MQTT lightweight publish-subscribe protocol. At the same time, independent communication topics are divided to isolate and transmit physiological data, action data, and device status data separately to avoid mutual interference between different types of data and improve the independence, security, and orderliness of data transmission.

[0059] The data packetization strategy and transmission frequency of the sensing module are optimized. Data frame compression technology is used to compress the length of single packet data, reduce redundant data transmission, reduce bandwidth usage, and improve the real-time performance and transmission efficiency of sensing data transmission.

[0060] This embodiment can effectively prevent the risks of unauthorized device access and identity impersonation; ensure that the data transmission process of physiological, action and device status is not stolen or tampered with, and achieve classified isolation and secure and reliable transmission; reduce the occupation of transmission redundancy bandwidth, and significantly improve the transmission efficiency and real-time performance of sensor data.

[0061] Example 5 Based on Embodiment 1, this embodiment provides a superior visual interface.

[0062] To facilitate martial arts judging and data management, in some preferred embodiments, the system of this invention further includes a data visualization and basic management module: connected to the collaborative quantitative evaluation module, it provides a web-based visualization interface, including: real-time refreshing and display of the athlete's physiological indicator curves, force curves, power radar charts, and total evaluation scores; achieving dynamic visualization of athletic status and skill level. In other preferred embodiments, a local data storage warehouse is built in the background to completely archive the original sensor data, feature sequences, various scores, and evaluation parameters of each competition or training process; categorized search conditions are designed, allowing for precise filtering by time, personnel, and session, and supporting dynamic playback of the complete data flow to reproduce the changes in indicators and evaluation results throughout the process; the system has a built-in report parsing and format conversion engine that can automatically summarize multi-dimensional evaluation data, indicator curve statistics, and comprehensive evaluation results, automatically formatting them according to standard templates to generate analysis reports in PDF / Excel formats, and enabling one-click export.

[0063] This embodiment can intuitively and in real time present multi-dimensional data on physiology, exertion, and comprehensive evaluation, with clear and easy-to-understand data display; it enables long-term data storage, rapid retrieval, and full playback, facilitating training technique review and status assessment; it supports one-click export of multi-format analysis reports, standardizing the output of evaluation results and improving the efficiency of data management in competitions and daily training.

[0064] Example 6 Based on the same technical concept, this embodiment also provides a computer device, such as... Figure 2 As shown, the computer device includes a memory 11 and a processor 12. The memory 11 stores a computer program. When the processor 12 runs the computer program stored in the memory 11, it realizes all the functions of the multimodal wearable system for quantitative evaluation of qi and strength coordination for martial arts training and competition.

[0065] Example 7 Based on the same technical concept, this embodiment also provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements all the functions of the multimodal wearable system for quantitative evaluation of qi and strength coordination for martial arts training and competition.

[0066] Experimental Example 1 To further verify the beneficial effects of the technical solution provided in this disclosure, the following experimental example is provided. This experimental example, through performance comparison with a control example, demonstrates that the technical solution of this invention has superior practical value in quantitative evaluation tasks for martial arts competitions and training.

[0067] 1. Experimental Objective This experiment aims to compare the performance of the embodiment of the present invention (which synchronously collects athlete's pulse signal and IMU sensor data through a sensing module; the collaborative quantitative evaluation module uses a sequence extraction unit to obtain heart rate variability sequence and movement force sequence respectively; then, the corresponding scores are calculated by each unit of strength coordination, force concentration, and movement stability through Pearson correlation coefficient, peak acceleration to standard deviation ratio, and Euclidean distance error; finally, the comprehensive evaluation unit weights and fuses the three scores to output the total evaluation score) with the following control example in a martial arts competition evaluation task: Comparative example: A scheme based on visual skeleton key point detection (Publication No.: CN118470750A).

[0068] 2. Experimental Setup Study Subjects: Eighteen healthy martial arts athletes were selected as experimental subjects, including 10 male athletes and 8 female athletes; aged 18-28 years, with an average age of 22.3±2.1 years; training years of 3-12 years, with an average training year of 6.5±2.4 years; all had no history of sports injuries, cardiovascular diseases, or neurological diseases, had not recently undergone high-intensity fatigue training, were in good physical condition, and were able to complete standard martial arts movements according to the experimental requirements. The martial arts movements in the experiment included three types: basic postures and stances: fist salute, preparatory stance, bow stance, horse stance, empty stance, and crouching stance; core punching techniques: punching, pushing palm, chopping punch, smashing punch, and downward punch; and main leg techniques: front kick, side kick, spring kick, horizontal kick, and stomping kick.

[0069] 3. Experimental Procedure The execution of the experimental example of the present invention: According to the method described in the aforementioned preferred embodiment, firstly, the martial arts athletes participating in the experiment were equipped with suitable sensing devices. Among them, a wearable smart ring integrating a multi-wavelength PPG photoplethysmography sensor was used as the core acquisition terminal. The IMU sensing unit was fixed to the athlete's torso and major force joints (6 key joints: wrist, elbow, shoulder, waist and hip, knee and ankle, with a single node weight ≤5g, fixed by elastic straps without restricting joint movement). All sensing devices were calibrated to ensure that the acquisition accuracy met the experimental requirements.

[0070] The collaborative quantitative evaluation module is activated to initialize each unit. The sampling frequency of the sequence extraction unit is set (to be consistent with the acquisition frequency of the sensing module to avoid data misalignment). The weighted fusion coefficient of the comprehensive evaluation unit is preset (based on the evaluation focus of martial arts competitions, the weights of the three indicators of strength coordination, force concentration, and movement stability are reasonably allocated; in this experiment, the weights are...). , , (Ensure that the weight allocation is scientific and traceable).

[0071] The system clearly defines the martial arts movements (such as fist techniques and weapon movements) that athletes need to perform, and sets the duration and number of repetitions for each movement. Pulse signals and IMU sensor data (including acceleration and angular velocity) are simultaneously collected from the athletes as they perform the preset martial arts movements. The timestamps of both types of data are synchronized and transmitted to the collaborative quantitative evaluation module for preprocessing. Then, the sequence extraction unit extracts heart rate variability sequences and movement force intensity sequences to provide data support for subsequent indicator calculations. Based on the extracted sequences, the three calculation units of the collaborative quantitative evaluation module complete the quantitative scoring of the corresponding indicators. The comprehensive evaluation unit calls the preset weighted fusion rules to perform weighted calculations on the scores of the three indicators, obtaining the athlete's total evaluation score for the martial arts movements. The weighting process strictly follows the preset weight allocation ratio to ensure that the scoring results comprehensively and objectively reflect the athlete's skill level. Finally, the system outputs the total evaluation score, and simultaneously outputs the scores and calculation process of the three individual indicators, providing complete evaluation data support for experimental comparative analysis.

[0072] The execution of the example: adopts a hybrid architecture of "edge GPU box + cloud asynchronous analysis", and its end-to-end latency calculation method is as follows: camera captures images → edge GPU skeleton key point detection → cloud action comparison and analysis → score result return.

[0073] 4. Experimental Results and Analysis The experimental results are shown in Table 1. This experiment mainly evaluates the performance of the present invention and the control example from three dimensions: real-time performance, robustness and deployment efficiency, with average end-to-end latency, environmental adaptability and system deployment time as the core indicators.

[0074] Table 1 Comparison of Experimental Results As shown in the table, this invention outperforms the control example in all three performance indicators. Firstly, it avoids the image transmission, cloud processing, and result feedback latency issues inherent in the control example's "edge GPU + cloud asynchronous analysis" approach. Therefore, the average end-to-end latency is only 187ms, significantly lower than the control example's 520ms, ensuring real-time feedback for martial arts competition evaluation. Secondly, this invention uses pulse and IMU sensor data for evaluation, unaffected by environmental factors such as light or occlusion. This overcomes the inherent sensitivity of visual skeletal keypoint detection schemes to scene conditions, providing a reliable data source for the stable quantification of core indicators such as energy coordination, force concentration, and movement stability. Finally, this invention only requires wearing a sensor device for deployment, eliminating the need for venue modifications or complex hardware infrastructure. System deployment time can be controlled within 10 minutes, significantly lowering the barrier to entry for competitions and training. In contrast, the control example, involving camera installation, edge device debugging, and cloud link configuration, requires over 60 minutes for deployment. This invention ensures that the final evaluation results comprehensively and objectively reflect the athlete's performance level.

[0075] 5. Experimental Conclusions This invention overcomes the inherent defects of existing visual evaluation schemes. Through collaborative quantitative evaluation of pulse and IMU sensor data, it achieves a comprehensive performance improvement in terms of low latency, high robustness, and rapid deployment. The output evaluation results are both comprehensive and objective, providing technical support for martial arts competition judging and technical analysis.

[0076] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A multimodal wearable system for quantitative assessment of Qi-strength coordination in martial arts training and competition, characterized in that, The system includes: Sensing module: used to simultaneously acquire the athlete's pulse signal and IMU sensor data; Collaborative Quantitative Evaluation Module: Connected to the sensing module, including: Sequence extraction unit: used to extract heart rate variability sequence and force exertion sequence based on the pulse signal and the IMU sensor data, respectively; Strength coordination acquisition unit: connected to the sequence extraction unit, used to align the heart rate variability sequence and the force exertion sequence, calculate the Pearson correlation coefficient, and obtain the strength coordination score; Force concentration acquisition unit: connected to the sequence extraction unit, used to calculate the ratio of the peak acceleration of each action to the standard deviation of the acceleration sequence based on the force intensity sequence of the action, and obtain the force concentration score; Action stability acquisition unit: connected to the sequence extraction unit, used to calculate the Euclidean distance error between the action force sequence and the preset action sequence, and obtain the action stability score; The comprehensive evaluation unit is connected to the force coordination acquisition unit, the force concentration acquisition unit, and the movement stability acquisition unit, respectively. It is used to weight and fuse the force coordination score, the force concentration score, and the movement stability score to obtain a total evaluation score; and is used to evaluate the athlete's movement.

2. The multimodal wearable system for quantitative evaluation of qi and strength coordination in martial arts training and competition, based on claim 1, is characterized in that... The system further includes a preprocessing module connected to the sensing module, and the preprocessing module performs preprocessing in the following ways: S1. Filter the pulse signal and the IMU sensing data respectively; and use the acceleration signal in the IMU sensing data as reference noise to eliminate motion interference in the pulse signal through an adaptive noise cancellation algorithm; to obtain the filtered pulse signal and IMU sensing data; S2. Outlier identification and removal are performed on the filtered pulse signal and the IMU sensing data; signal completion is performed using linear interpolation and spline fitting methods to obtain the completed pulse signal and IMU sensing data; S3. Extract the heart rate variability sequence and the force exertion sequence based on the completed pulse signal and the IMU sensor data, respectively; and perform time synchronization of the heart rate variability sequence and the force exertion sequence through a preset time synchronization mechanism.

3. The multimodal wearable system for quantitative evaluation of qi and strength coordination in martial arts training and competition, based on claim 2, is characterized in that... The preset timing synchronization mechanism includes: S301. Based on the NTP network time protocol, complete the clock reference calibration between the sensing module and the gateway, and output a unified clock source; S302. Based on the unified clock source, add timestamps to the heart rate variability sequence and the force exertion sequence, and output the heart rate variability sequence and the force exertion sequence with timestamps; S303. The dynamic time warping algorithm is used to align the timestamped heart rate variability sequence and the force exertion sequence.

4. The multimodal wearable system for quantitative evaluation of qi and strength coordination in martial arts training and competition, based on claim 1, is characterized in that... The system also includes a sensor access and transmission module connected to the sensing module. The implementation method of the sensor access and transmission module includes: pre-setting a unique hardware ID and key for each sensing unit in the sensing module, and the gateway having a built-in hardware encryption unit and device whitelist; when the sensing unit is accessed, the gateway verifies the hardware ID and compares the key, and then assigns a temporary session ID, which will expire after the end of a single event. The data collected by the sensing module is encrypted using the national cryptographic SM4 algorithm and transmitted to the gateway based on the MQTT lightweight publish-subscribe protocol. Physiological data, action data, and device status data are classified and transmitted separately through independent topics. The data packetization strategy and transmission frequency of the sensing module are optimized, and data frame compression technology is used to compress the length of a single packet of data.

5. The multimodal wearable system for quantitative evaluation of qi and strength coordination in martial arts training and competition as described in claim 1, characterized in that, The sensing module includes: Physiological sensing unit: used to collect data including the athlete's heart rate, blood oxygen saturation, pulse wave velocity, and heart rate variability. Motion biomechanics sensing unit: used to collect data including the athlete's motion acceleration, angular velocity, and posture angle.

6. The multimodal wearable system for quantitative evaluation of qi and strength coordination in martial arts training and competition, based on claim 1, is characterized in that... The system also includes a data visualization and basic management module: connected to the collaborative quantitative evaluation module, it provides a web-based visualization interface, including: real-time display of athletes' physiological indicator curves, force curves, power radar charts, and total evaluation scores; it supports local storage, retrieval, and playback of competition data, as well as export of analysis reports in PDF / Excel format.

7. A computer device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to achieve the function of the multimodal wearable system for quantitative evaluation of qi and strength coordination for martial arts training and competition, as described in any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for enabling the computer to perform the functions of the multimodal wearable system for quantitative evaluation of qi and strength coordination for martial arts training and competition as described in any one of claims 1 to 6.

Images