An artificial intelligence sports injury early warning system and method

Abstract

This invention relates to the field of sports injury early warning technology, specifically to an artificial intelligence sports injury early warning system and method. The method includes: establishing a basic identity and health record for athletes; simultaneously collecting multimodal raw data including biomechanical data, physiological load data, and environmental parameter data during the athlete's exercise; processing the multimodal raw data; matching and comparing the feature dataset with a database of injury-prone behavior feature parameters; generating graded early warning information based on the risk probability interval when the probability of injury exceeds a preset dynamic threshold; and generating corresponding personalized intervention plans based on the graded early warning information. The system includes: an information filing module, a data acquisition module, a data preprocessing module, a risk assessment and early warning module, and an intervention feedback optimization module. This system enables the comprehensive capture of complex causes of sports injuries, improving the comprehensiveness and accuracy of injury risk assessment.

Description

Technical Field

[0001] This invention relates to the field of sports injury early warning technology, and in particular to an artificial intelligence sports injury early warning system and method. Background Technology

[0002] Sports injuries are a core limiting factor affecting the health, athletic performance, and training careers of athletes. Especially in professional competitive sports, mass fitness, and rehabilitation training, sports injuries not only lead to short-term training interruptions and a decline in competitive levels, but can also cause chronic strain, resulting in irreversible effects on long-term athletic development. Traditional sports injury prevention and control mainly rely on the experience and judgment of coaches and sports medicine experts, as well as athletes' subjective feedback on physical discomfort. This approach is inherently subjective and delayed: experience-based judgment is easily influenced by differences in the professional level and cognitive abilities of personnel, making it difficult to accurately identify early signs of injury; subjective feedback is often only noticed when the injury has progressed to a stage with obvious symptoms, missing the optimal opportunity for early intervention.

[0003] In existing technologies, heart rate data is collected solely through heart rate sensors, and heart rate changes are used as the basis for determining exercise load and injury risk. This results in a single data dimension, making it difficult to comprehensively characterize the causes of injury. Summary of the Invention

[0004] The purpose of this invention is to provide an artificial intelligence sports injury early warning system and method, which aims to solve the technical problem in the prior art that only collects heart rate data through a heart rate sensor and uses heart rate changes as the basis for judging exercise load and injury risk, resulting in a single data dimension and difficulty in comprehensively characterizing the causes of injury.

[0005] To achieve the above objectives, the present invention employs an artificial intelligence-based sports injury early warning method, comprising the following steps: Obtain athletes' basic information, assign unique numbers and codes to the basic information, and enter it into the information database to establish athletes' basic identity and health records; Equip athletes with sensing and visual acquisition devices and complete the wearing process to simultaneously collect multimodal raw data, including biomechanical data, physiological load data, and environmental parameter data, during the athlete's exercise. Data cleaning, noise suppression, and spatiotemporal synchronization alignment are performed sequentially on the multimodal raw data to generate a standardized feature dataset; A database of injury-prone behavior features for different sports is constructed. The standardized feature dataset is matched and compared with the database of injury-prone behavior features to calculate the probability of injury risk. When the probability of injury risk is greater than a preset dynamic threshold, a graded early warning information is generated based on the risk probability interval. Based on the tiered early warning information, a corresponding personalized intervention plan is generated and pushed to the athlete's target terminal; at the same time, feedback data on the athlete's implementation of the intervention plan is received to optimize and iterate the strategy.

[0006] Among the steps involved in obtaining athletes' basic information, assigning unique numbers and codes to this information, and entering it into the information database to establish athletes' basic identity and health records: Collect basic information on athletes, including their identification, sport, age, gender, body shape indicators, history of injury, daily training / competition plans, and health check-up reports. Classify and organize this information according to identity attributes, sport attributes, and health attributes, and remove duplicate and invalid information items. A unique number is generated for each piece of basic information by using a combination rule of athlete identification number, collection timestamp, and information category, and an associated index is established for the number. The completeness of the coded basic information is verified, and missing items are checked and supplemented through communication with athletes. The verified and completed basic information is entered into the information database in a unified format, and a basic identity and health record exclusive to the athlete is generated simultaneously, with associated numbers and information items in the record.

[0007] Among these steps, the process of equipping athletes with sensing and visual acquisition devices and completing their donning, and simultaneously collecting multimodal raw data including biomechanical data, physiological load data, and environmental parameter data during the athlete's exercise includes: Based on the athlete's sport, physical characteristics, and data collection needs, select and configure sensing and visual acquisition equipment. Based on the characteristics of the sport, set the acquisition frequency, data transmission format, and synchronization trigger conditions for each device; The data acquisition equipment is used to simultaneously collect biomechanical data, physiological load data, and environmental parameter data during athletes' daily training and competitions, and the raw data is stored in real time.

[0008] Among them, the step of selecting and configuring sensing and visual acquisition equipment according to the athlete's sport, physical characteristics, and data collection needs is as follows: The sensing devices include inertial measurement units, flexible pressure sensors, and electromyography sensors for capturing biomechanical data, as well as heart rate sensors and bioelectrical impedance sensors for acquiring physiological load data. The visual acquisition device uses an infrared camera.

[0009] Among these steps, the process involves using data acquisition equipment to simultaneously collect biomechanical data, physiological load data, and environmental parameter data during athletes' daily training and competitions, and storing the collected raw data in real time: Biomechanical data include joint angles, acceleration, angular velocity, plantar pressure distribution, and electromyographic signals on the muscle surface; physiological load data include heart rate, heart rate variability, muscle impedance, and phase angle; environmental parameter data include site temperature and humidity, coefficient of friction, and light intensity.

[0010] Among them, in the steps of sequentially performing data cleaning, noise suppression, and spatiotemporal synchronization alignment on the multimodal raw data to generate a standardized feature dataset: Remove missing values, outliers, and duplicate records from the original multimodal data; For biomechanical and physiological load data, Gaussian filtering is used to suppress noise and filter out invalid noise signals caused by equipment vibration and environmental interference during the acquisition process. Based on the timestamps of the visual acquisition device and the sampling timestamps of the sensor device, a linear interpolation method is used to complete the spatiotemporal alignment of multimodal data, ensuring accurate matching of biomechanical, physiological load and environmental parameter data at the same time point; The aligned clean data is subjected to feature engineering to extract injury-related features, and the extracted features are normalized to generate a standardized feature dataset with uniform format and consistent dimensions; the injury-related features include joint range of motion, muscle exertion efficiency, and acute-to-chronic load ratio.

[0011] Among these steps, the process involves pre-constructing a database of injury-prone behavioral characteristic parameters for different sports, matching and comparing a standardized feature dataset with the database to calculate the probability of injury risk, and generating tiered early warning information based on the risk probability interval when the probability of injury risk exceeds a preset dynamic threshold. By statistical analysis of historical injury cases in different sports, combined with professional analysis in sports medicine, and integrating a large amount of athlete training / competition data, we sorted out the characteristics of high-risk movements, load thresholds, and environmental triggers for each sport, established a database of injury-prone behavior characteristics by sports, and labeled the risk weight of each characteristic. Based on the athlete's sport, history of injury, real-time physiological state and training stage, a basic risk threshold is preset, and the threshold is dynamically adjusted for different scenarios. For high-intensity sports and athletes with a history of injury, the threshold is appropriately lowered to improve the sensitivity of the warning. The standardized feature dataset is matched and compared with the injury-prone behavior feature parameter library of the corresponding sports item by item. The athlete's current injury risk probability is calculated based on the feature matching degree, risk weight and feature occurrence frequency. The probability of damage risk is divided into low-risk, medium-risk, high-risk, and extremely high-risk zones. When the calculated probability of risk is greater than the dynamic threshold, a graded early warning information containing risk level, risk cause, and high-risk location is generated according to the corresponding interval.

[0012] Among the steps, the process involves matching and generating corresponding personalized intervention plans based on tiered early warning information and pushing them to the athlete's target terminal; simultaneously receiving feedback data from the athlete on the implementation of the intervention plan and optimizing and iterating the strategy accordingly. Based on the risk level of the graded early warning information, corresponding personalized intervention plans are matched. For low-risk cases, suggestions for optimizing exercise posture and mild stretching training are pushed; for medium-risk cases, suggestions for adjusting exercise intensity and local muscle relaxation training are pushed; for high-risk cases, suggestions for pausing exercise, resting and observing, and making appointments for rehabilitation therapy are pushed; and for extremely high-risk cases, instructions for emergency medical treatment and location sharing are pushed. The matched intervention plan will be pushed to the athlete's mobile terminal at a specified frequency, along with the warning information and the association between the intervention plan and the warning information. Receive feedback data from athletes on the implementation of the intervention program, including whether the intervention program was implemented, changes in physical condition after the intervention, and subsequent exercise data; The collected feedback data is analyzed, and combined with the athlete's historical data in the information database, the feature weights of the injury-prone behavior characteristic parameter database are optimized, the dynamic thresholds are adjusted, and the adaptation strategies of the intervention program are optimized.

[0013] This invention also provides an artificial intelligence sports injury early warning system, comprising an information filing module, a data acquisition module, a data preprocessing module, a risk assessment and early warning module, and an intervention feedback optimization module; wherein: The information filing module is used to obtain the athlete's basic information, assign a unique number to the basic information, and enter it into the information database to establish the athlete's basic identity and health record. The data acquisition module is used to collect biomechanical data, physiological load data, and environmental parameter data during the athlete's exercise process; The data preprocessing module is used to sequentially perform data cleaning, noise suppression, and spatiotemporal synchronization alignment on the multimodal raw data to generate a standardized feature dataset; The risk assessment and early warning module is used to construct a database of injury-prone behavior characteristic parameters for different sports, and to match and compare the standardized feature dataset with the database of injury-prone behavior characteristic parameters to calculate the probability of injury risk. The intervention feedback optimization module is used to match and generate corresponding personalized intervention plans based on the graded early warning information and push them to the athlete's target terminal; at the same time, it receives the athlete's execution feedback data on the intervention plan and optimizes and iterates the strategy.

[0014] This invention discloses an artificial intelligence sports injury early warning system and method, which employs the information filing module, data acquisition module, data preprocessing module, risk assessment and early warning module, and intervention feedback optimization module to perform the following steps: acquiring the athlete's basic information, uniquely numbering and encoding the basic information, and recording it in an information database to establish the athlete's basic identity and health record; equipping the athlete with sensing devices and visual acquisition devices and completing their donning, simultaneously collecting multimodal raw data including biomechanical data, physiological load data, and environmental parameter data during the athlete's exercise; and sequentially performing data cleaning, noise suppression, and spatiotemporal synchronization alignment processing on the multimodal raw data. A standardized feature dataset is generated; a pre-constructed database of injury-prone behavior feature parameters for different sports is built, and the standardized feature dataset is matched and compared with the database to calculate the probability of injury risk; when the probability of injury risk exceeds a preset dynamic threshold, a graded early warning information is generated based on the risk probability interval; based on the graded early warning information, a corresponding personalized intervention plan is generated and pushed to the athlete's target terminal; at the same time, feedback data on the athlete's implementation of the intervention plan is received to optimize and iterate the strategy; through the above methods, the complex causes of sports injuries can be comprehensively captured, and the causes of sports injuries can be represented in a full-dimensional and three-dimensional way, improving the comprehensiveness and accuracy of injury risk assessment. Attached Figure Description

[0015] 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.

[0016] Figure 1 This is a flowchart of the steps of the artificial intelligence sports injury early warning method of the present invention.

[0017] Figure 2 This is a flowchart of steps S100 of the present invention.

[0018] Figure 3 This is a flowchart of steps S200 of the present invention.

[0019] Figure 4 This is a flowchart of steps S300 of the present invention.

[0020] Figure 5 This is a flowchart of steps S400 of the present invention.

[0021] Figure 6 This is a flowchart of steps S500 of the present invention.

[0022] Figure 7 This is a schematic diagram of the principle of the artificial intelligence sports injury early warning system of the present invention.

[0023] 601-Information Filing Module, 602-Data Acquisition Module, 603-Data Preprocessing Module, 604-Risk Assessment and Early Warning Module, 605-Intervention Feedback and Optimization Module. Detailed Implementation

[0024] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0025] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0026] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0027] Please see Figures 1-6 This invention provides an artificial intelligence-based sports injury early warning method, comprising the following steps: S100: Obtain the athlete's basic information, assign a unique number to the basic information, and enter it into the information database to establish the athlete's basic identity and health record.

[0028] In this embodiment, the athlete's basic information is obtained, uniquely coded, and entered into the information database to establish the athlete's basic identity and health record. The specific process is as follows: S101: Collect basic information on athletes, including their identity, sport, age, gender, body shape indicators, history of past injuries, daily training / competition plans, and health check-up reports. Classify and organize the information according to identity attributes, sport attributes, and health attributes, and remove duplicate and invalid information items. S102: Using a combination rule of athlete identification number, collection timestamp, and information category, a unique number code is generated for each piece of basic information, and an associated index is established for the number; S103: Perform integrity verification on the coded basic information, check for missing items and complete them through communication with athletes; S104: Enter the verified and completed basic information into the information database in a unified format, and simultaneously generate a basic identity and health record exclusive to the athlete, with associated numbers and information items in the record.

[0029] In the above process, the basic information of athletes is first collected, including their identity, sport, age, gender, body shape indicators, history of injury, daily training / competition plans, and health examination reports. This information is then categorized and organized according to identity attributes, sport attributes, and health attributes, and duplicate and invalid information items are removed. Next, a unique number is generated for each piece of basic information using a combination rule of athlete identification number, collection timestamp, and information category, and an associated index is established for the number. The completeness of the coded basic information is verified, and missing items are checked and supplemented through communication with the athlete. Subsequently, the verified and supplemented basic information is entered into the information database in a unified format, and a unique basic identity and health file for each athlete is generated simultaneously, with the associated number and information items in the file.

[0030] S200: Equips athletes with sensing and visual acquisition devices and completes the wearing process. During the athlete's exercise, it simultaneously collects multimodal raw data, including biomechanical data, physiological load data, and environmental parameter data.

[0031] In this embodiment, the athlete is equipped with sensing and visual acquisition devices and wears them. During the athlete's exercise, multimodal raw data, including biomechanical data, physiological load data, and environmental parameter data, are collected simultaneously. The specific process is as follows: S201: Select and equip sensing and visual acquisition devices according to the athlete's sport, physical characteristics and data collection needs; S202: Based on the characteristics of the sport, set the acquisition frequency, data transmission format and synchronization trigger conditions for each device; S203: Using data acquisition equipment, biomechanical data, physiological load data, and environmental parameter data are collected simultaneously during athletes' daily training and competitions, and the raw data is stored in real time.

[0032] In the above process, sensing and visual acquisition devices are selected based on the athlete's sport, physical characteristics, and data acquisition needs. The sensing devices include inertial measurement units, flexible pressure sensors, and electromyography (EMG) sensors for capturing biomechanical data, and heart rate sensors and bioelectrical impedance sensors for collecting physiological load data. Infrared cameras are used for visual acquisition. The acquisition frequency, data transmission format, and synchronization triggering conditions of each device are set according to the characteristics of the sport. The acquisition devices are then used to simultaneously collect biomechanical data, physiological load data, and environmental parameter data during the athlete's daily training and competition, storing the raw data in real time. Biomechanical data includes joint angles, acceleration, angular velocity, plantar pressure distribution, and muscle surface EMG signals; physiological load data includes heart rate, heart rate variability, muscle impedance, and phase angle; and environmental parameter data includes venue temperature and humidity, coefficient of friction, and light intensity.

[0033] S300: Performs data cleaning, noise suppression, and spatiotemporal synchronization alignment on the multimodal raw data in sequence to generate a standardized feature dataset.

[0034] In this embodiment, data cleaning, noise suppression, and spatiotemporal synchronization alignment are sequentially performed on the multimodal raw data to generate a standardized feature dataset. The specific process is as follows: S301: Remove missing values, outliers, and duplicate records from the original multimodal data; S302: For biomechanical data and physiological load data, Gaussian filtering is used to suppress noise and filter out invalid noise signals generated by equipment vibration and environmental interference during the acquisition process. S303: Based on the timestamps of the visual acquisition device and the sampling timestamps of the sensing device, the spatiotemporal alignment of multimodal data is completed by linear interpolation to ensure accurate matching of biomechanical, physiological load and environmental parameter data at the same time point; S304: Perform feature engineering on the aligned clean data to extract injury-related features, and normalize the extracted features to generate a standardized feature dataset with uniform format and consistent dimensions; among which, injury-related features include joint range of motion, muscle exertion efficiency, and acute-to-chronic load ratio.

[0035] In the above process, missing values, outliers, and duplicate records in the original multimodal data are removed. Gaussian filtering is used to suppress noise in the biomechanical and physiological load data, filtering out invalid noise signals caused by equipment vibration and environmental interference during the acquisition process. Then, based on the timestamps of the visual acquisition device and the sampling timestamps of the sensor device, linear interpolation is used to complete the spatiotemporal alignment of the multimodal data, ensuring accurate matching of biomechanical, physiological load, and environmental parameter data at the same time point. Subsequently, feature engineering is performed on the aligned clean data to extract injury-related features, and the extracted features are normalized to generate a standardized feature dataset with a unified format and consistent dimensions. The injury-related features include joint range of motion, muscle exertion efficiency, and the ratio of acute to chronic load.

[0036] S400: Construct a vulnerability behavior feature parameter library for different sports, match and compare the standardized feature dataset with the vulnerability behavior feature parameter library, and calculate the probability of injury risk; when the probability of injury risk is greater than the preset dynamic threshold, generate graded early warning information based on the risk probability interval.

[0037] In this embodiment, a vulnerability behavior feature parameter library for different sports is constructed. The standardized feature dataset is matched and compared with the vulnerability behavior feature parameter library to calculate the injury risk probability. When the injury risk probability exceeds a preset dynamic threshold, a graded early warning information is generated based on the risk probability interval. The specific process is as follows: S401: Statistically analyze historical injury cases of different sports, combine sports medicine professional analysis, integrate a large amount of athlete training / competition data, sort out the characteristics of high-risk movements, load thresholds and environmental triggers of each sport, establish a database of injury-prone behavior characteristic parameters by sports, and label the risk weight of each characteristic. S402: Based on the athlete's sport, history of injury, real-time physiological state and training stage, a basic risk threshold is preset, and the threshold is dynamically adjusted for different scenarios. For high-intensity sports and athletes with a history of injury, the threshold is appropriately lowered to improve the sensitivity of the warning. S403: Match and compare the standardized feature dataset with the injury-prone behavior feature parameter library of the corresponding sport item by item, and calculate the athlete's current injury risk probability based on feature matching degree, risk weight and feature occurrence frequency. S404: Divide the probability of damage risk into low-risk, medium-risk, high-risk, and extremely high-risk zones. When the calculated probability of risk is greater than the dynamic threshold, generate graded early warning information containing risk level, risk cause, and high-risk location according to the corresponding interval.

[0038] In the above process, firstly, historical injury cases of different sports are statistically analyzed, combined with professional analysis of sports medicine, and a large amount of athlete training / competition data are integrated to sort out the characteristics of high-risk movements, load thresholds, and environmental triggering characteristics of each sport. A vulnerability behavior characteristic parameter library is established according to sports, and the risk weight of each feature is marked. Then, based on the athlete's sport, past injury history, real-time physiological state, and training stage, a basic risk threshold is preset, and the threshold is dynamically adjusted for different scenarios. For high-intensity sports and athletes with a history of injury, the threshold is appropriately lowered to improve the sensitivity of the warning. Next, the standardized feature dataset is matched and compared item by item with the vulnerability behavior characteristic parameter library of the corresponding sports. Based on the feature matching degree, risk weight, and feature occurrence frequency, the athlete's current injury risk probability is comprehensively calculated. The injury risk probability is divided into low-risk, medium-risk, high-risk, and extremely high-risk zones. When the calculated risk probability is greater than the dynamic threshold, a graded warning information containing risk level, risk cause, and high-risk body part is generated according to the corresponding interval.

[0039] S500: Based on tiered early warning information, it matches and generates corresponding personalized intervention plans and pushes them to the athlete's target terminal; at the same time, it receives feedback data from the athlete on the implementation of the intervention plan and optimizes and iterates the strategy.

[0040] In this embodiment, based on tiered early warning information, a corresponding personalized intervention plan is generated and pushed to the athlete's target terminal; simultaneously, feedback data from the athlete on the implementation of the intervention plan is received, and the strategy is optimized and iterated. The specific process is as follows: S501: Based on the risk level of the graded early warning information, a corresponding personalized intervention plan is matched. For low risk, suggestions for optimizing exercise posture and mild stretching training are pushed; for medium risk, suggestions for adjusting exercise intensity and local muscle relaxation training are pushed; for high risk, suggestions for pausing exercise, resting and observing, and making appointments for rehabilitation therapy are pushed; and for very high risk, instructions for emergency medical treatment and location sharing are pushed. S502: Push the matched intervention plan to the athlete's mobile terminal at a specified frequency, and at the same time push the warning information and the association description of the intervention plan; S503: Receive feedback data from athletes on the implementation of the intervention program, including whether the intervention program was implemented, changes in physical condition after the intervention, and subsequent exercise data; S504: Analyze the collected feedback data, combine it with the athlete's historical data in the information database, optimize the feature weights of the injury-prone behavior characteristic parameter database, adjust the dynamic thresholds, and adapt the intervention program to the appropriate strategies.

[0041] In the above process, firstly, based on the risk level of the graded early warning information, corresponding personalized intervention plans are matched. For low-risk cases, suggestions for optimizing exercise posture and mild stretching training are pushed; for medium-risk cases, suggestions for adjusting exercise intensity and local muscle relaxation training are pushed; for high-risk cases, suggestions for pausing exercise, resting and observing, and scheduling rehabilitation therapy are pushed; and for extremely high-risk cases, instructions for emergency medical treatment and location sharing are pushed. Then, the matched intervention plans are pushed to the athletes' mobile terminals at a specified frequency, along with an explanation of the association between the early warning information and the intervention plan. Feedback data on the athletes' implementation of the intervention plans is received, including whether the intervention plan was implemented, changes in physical condition after the intervention, and subsequent exercise data. The collected feedback data is analyzed and combined with the athletes' historical data in the information database to optimize the feature weights of the injury-prone behavior characteristic parameter library, adjust dynamic thresholds, and adapt the intervention plan to the appropriate strategies.

[0042] Please see Figure 7 This invention provides an artificial intelligence-based sports injury early warning system, comprising an information filing module, a data acquisition module, a data preprocessing module, a risk assessment and early warning module, and an intervention feedback optimization module; wherein: The information filing module is used to obtain the athlete's basic information, assign a unique number to the basic information, and enter it into the information database to establish the athlete's basic identity and health record. The data acquisition module is used to collect biomechanical data, physiological load data, and environmental parameter data during the athlete's exercise process; The data preprocessing module is used to sequentially perform data cleaning, noise suppression, and spatiotemporal synchronization alignment on the multimodal raw data to generate a standardized feature dataset; The risk assessment and early warning module is used to construct a database of injury-prone behavior characteristic parameters for different sports, and to match and compare the standardized feature dataset with the database of injury-prone behavior characteristic parameters to calculate the probability of injury risk. The intervention feedback optimization module is used to match and generate corresponding personalized intervention plans based on the graded early warning information and push them to the athlete's target terminal; at the same time, it receives the athlete's execution feedback data on the intervention plan and optimizes and iterates the strategy.

[0043] In this embodiment, the basic information of athletes is obtained through the information filing module, which assigns a unique number to the basic information and enters it into the information database to establish the athletes' basic identity and health records. The data acquisition module collects biomechanical data, physiological load data, and environmental parameter data during the athletes' exercise process. The data preprocessing module performs data cleaning, noise suppression, and spatiotemporal synchronization alignment on the multimodal raw data in sequence to generate a standardized feature dataset. The risk assessment and early warning module constructs a database of injury-prone behavior feature parameters for different sports, matches and compares the standardized feature dataset with the database, and calculates the probability of injury risk. The intervention feedback optimization module generates corresponding personalized intervention plans based on the graded early warning information and pushes them to the athletes' target terminals. At the same time, it receives feedback data from athletes on the implementation of the intervention plan and optimizes and iterates the strategy.

[0044] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

[0045] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. An artificial intelligence-based sports injury early warning method, characterized in that, Includes the following steps: Obtain athletes' basic information, assign unique numbers and codes to the basic information, and enter it into the information database to establish athletes' basic identity and health records; Equip athletes with sensing and visual acquisition devices and complete the wearing process to simultaneously collect multimodal raw data, including biomechanical data, physiological load data, and environmental parameter data, during the athlete's exercise. Data cleaning, noise suppression, and spatiotemporal synchronization alignment are performed sequentially on the multimodal raw data to generate a standardized feature dataset; A database of injury-prone behavior characteristics is constructed for different sports. The standardized feature dataset is matched and compared with the database of injury-prone behavior characteristics to calculate the probability of injury risk. When the probability of damage risk exceeds a preset dynamic threshold, a graded early warning message is generated based on the risk probability range. Based on the tiered early warning information, a corresponding personalized intervention plan is generated and pushed to the athlete's target terminal; Simultaneously, feedback data from athletes on the implementation of the intervention plan is received to optimize and iterate the strategy.

2. The artificial intelligence sports injury early warning method as described in claim 1, characterized in that, In the steps of obtaining athletes' basic information, assigning unique numbers and codes to the basic information, and entering it into the information database to establish athletes' basic identity and health records: Collect basic information on athletes, including their identification, sport, age, gender, body shape indicators, history of injury, daily training / competition plans, and health check-up reports. Classify and organize this information according to identity attributes, sport attributes, and health attributes, and remove duplicate and invalid information items. A unique number is generated for each piece of basic information by using a combination rule of athlete identification number, collection timestamp, and information category, and an associated index is established for the number. The completeness of the coded basic information is verified, and missing items are checked and supplemented through communication with athletes. The verified and completed basic information is entered into the information database in a unified format, and a basic identity and health record exclusive to the athlete is generated simultaneously, with associated numbers and information items in the record.

3. The artificial intelligence sports injury early warning method as described in claim 1, characterized in that, In the process of equipping athletes with sensing and visual acquisition devices and completing the wearing process, and simultaneously collecting multimodal raw data including biomechanical data, physiological load data, and environmental parameter data during the athlete's exercise: Based on the athlete's sport, physical characteristics, and data collection needs, select and configure sensing and visual acquisition equipment. Based on the characteristics of the sport, set the acquisition frequency, data transmission format, and synchronization trigger conditions for each device; The data acquisition equipment is used to simultaneously collect biomechanical data, physiological load data, and environmental parameter data during athletes' daily training and competitions, and the raw data is stored in real time.

4. The artificial intelligence sports injury early warning method as described in claim 3, characterized in that, In the process of selecting and configuring sensing and visual acquisition devices based on the athlete's sport, physical characteristics, and data collection needs: The sensing devices include inertial measurement units, flexible pressure sensors, and electromyography sensors for capturing biomechanical data, as well as heart rate sensors and bioelectrical impedance sensors for acquiring physiological load data. The visual acquisition device uses an infrared camera.

5. The artificial intelligence sports injury early warning method as described in claim 4, characterized in that, In the process of using data acquisition equipment to simultaneously collect biomechanical data, physiological load data, and environmental parameter data during athletes' daily training and competitions, and storing the collected raw data in real time: Biomechanical data include joint angles, acceleration, angular velocity, plantar pressure distribution, and electromyographic signals on the muscle surface; physiological load data include heart rate, heart rate variability, muscle impedance, and phase angle; environmental parameter data include site temperature and humidity, coefficient of friction, and light intensity.

6. The artificial intelligence sports injury early warning method as described in claim 1, characterized in that, In the steps of sequentially performing data cleaning, noise suppression, and spatiotemporal synchronization alignment on the multimodal raw data to generate a standardized feature dataset: Remove missing values, outliers, and duplicate records from the original multimodal data; For biomechanical and physiological load data, Gaussian filtering is used to suppress noise and filter out invalid noise signals caused by equipment vibration and environmental interference during the acquisition process. Based on the timestamps of the visual acquisition device and the sampling timestamps of the sensor device, a linear interpolation method is used to complete the spatiotemporal alignment of multimodal data, ensuring accurate matching of biomechanical, physiological load and environmental parameter data at the same time point; The aligned clean data is subjected to feature engineering to extract damage-related features, and the extracted features are normalized to generate a standardized feature dataset with uniform format and consistent dimensions. The injury-related features include joint range of motion, muscle exertion efficiency, and acute-to-chronic load ratio.

7. The artificial intelligence sports injury early warning method as described in claim 1, characterized in that, In a pre-built database of injury-prone behavioral feature parameters for different sports, the standardized feature dataset is matched and compared with the database to calculate the probability of injury risk. In the step of generating graded early warning information based on the risk probability interval when the probability of damage is greater than the preset dynamic threshold: By statistical analysis of historical injury cases in different sports, combined with professional analysis in sports medicine, and integrating a large amount of athlete training / competition data, we sorted out the characteristics of high-risk movements, load thresholds, and environmental triggers for each sport, established a database of injury-prone behavior characteristics by sports, and labeled the risk weight of each characteristic. Based on the athlete's sport, history of injury, real-time physiological state and training stage, a basic risk threshold is preset, and the threshold is dynamically adjusted for different scenarios. For high-intensity sports and athletes with a history of injury, the threshold is appropriately lowered to improve the sensitivity of the warning. The standardized feature dataset is matched and compared item by item with the injury-prone behavior feature parameter library of the corresponding sports. The athlete's current injury risk probability is calculated based on the feature matching degree, risk weight and feature occurrence frequency. The probability of damage risk is divided into low-risk, medium-risk, high-risk, and extremely high-risk zones. When the calculated risk probability is greater than the dynamic threshold, a graded early warning information containing risk level, risk cause, and high-risk location is generated according to the corresponding interval.

8. The artificial intelligence sports injury early warning method as described in claim 1, characterized in that, In the steps of matching and generating corresponding personalized intervention plans based on tiered early warning information and pushing them to the athlete's target terminal; and simultaneously receiving feedback data from the athlete on the implementation of the intervention plan and optimizing and iterating the strategy: Based on the risk level of the graded early warning information, corresponding personalized intervention plans are matched. For low-risk cases, suggestions for optimizing exercise posture and mild stretching training are pushed; for medium-risk cases, suggestions for adjusting exercise intensity and local muscle relaxation training are pushed; for high-risk cases, suggestions for pausing exercise, resting and observing, and making appointments for rehabilitation therapy are pushed; and for extremely high-risk cases, instructions for emergency medical treatment and location sharing are pushed. The matched intervention plan will be pushed to the athlete's mobile terminal at a specified frequency, along with the warning information and the association between the intervention plan and the warning information. Receive feedback data from athletes on the implementation of the intervention program, including whether the intervention program was implemented, changes in physical condition after the intervention, and subsequent exercise data; The collected feedback data is analyzed, and combined with the athlete's historical data in the information database, the feature weights of the injury-prone behavior characteristic parameter database are optimized, the dynamic thresholds are adjusted, and the adaptation strategies of the intervention program are optimized.

9. An artificial intelligence sports injury early warning system, applied to the artificial intelligence sports injury early warning method as described in claim 1, characterized in that, It includes modules for information filing, data collection, data preprocessing, risk assessment and early warning, and intervention feedback and optimization; among which: The information filing module is used to obtain the athlete's basic information, assign a unique number to the basic information, and enter it into the information database to establish the athlete's basic identity and health record. The data acquisition module is used to collect biomechanical data, physiological load data, and environmental parameter data during the athlete's exercise process; The data preprocessing module is used to sequentially perform data cleaning, noise suppression, and spatiotemporal synchronization alignment on the multimodal raw data to generate a standardized feature dataset; The risk assessment and early warning module is used to construct a database of injury-prone behavior characteristic parameters for different sports, and to match and compare the standardized feature dataset with the database of injury-prone behavior characteristic parameters to calculate the probability of injury risk. The intervention feedback optimization module is used to match and generate corresponding personalized intervention plans based on the graded early warning information and push them to the athlete's target terminal; at the same time, it receives the athlete's execution feedback data on the intervention plan and optimizes and iterates the strategy.

Images