Real-time assessment and early warning system for sports injury risk

Abstract

The invention discloses a sports injury risk real-time assessment and early warning system. The system comprises an intelligent wearable device for collecting multi-dimensional physiological data and a remote analysis server for deep analysis; the server performs real-time risk assessment and graded early warning according to a personalized threshold value by fusing and analyzing the heart load, the body fluid balance state and muscle and joint stability characteristics extracted from kinematics parameters, and issues a specific intervention instruction; according to the invention, real-time perception of multiple risk factors of sports injury is realized through an end-cloud collaborative architecture, professional laboratory-level analysis ability is integrated into daily wearable equipment, the contradiction between single early warning dimension of existing equipment and incapability of real-time convenient application of a professional system is solved, accurate and prospective injury prevention guidance can be provided for a user, and the user experience is improved. And the safety of movement is greatly improved.

Description

Technical Field

[0001] This invention relates to the field of sports health monitoring technology, and in particular to a real-time assessment and early warning system for sports injury risk. Background Technology

[0002] In existing technologies, intelligent monitoring devices can typically track users' basic physiological and activity indicators such as heart rate, steps, and calorie consumption. Some advanced products can also provide single threshold alerts for excessive heart rate or excessive exercise. Some professional sports analysis systems, by combining multiple inertial sensors or surface electromyography sensors, can perform relatively complex offline analyses of athletes' technical movements and muscle loads in laboratory settings or specific training scenarios to assess their injury risk.

[0003] Most existing consumer-grade wearable devices have limited functionality, typically only recording and providing simple alerts for isolated, post-event statistical data on individual indicators such as heart rate. They lack the ability to comprehensively and in real-time perceive and assess various potential risk factors during exercise, such as dehydration, muscle fatigue, and movement abnormalities. For example, a runner might experience fluid imbalance due to continuous sweating (early dehydration) or decreased running stability due to fatigue, even when their heart rate is within a "safe" range. In such cases, a single heart rate monitor would completely ignore these important injury risk signals. Furthermore, while professional laboratory analysis systems can provide deeper insights, they rely on complex and expensive equipment, require professional operation and interpretation, and their results are significantly delayed, failing to provide immediate feedback and intervention during exercise. Therefore, they are difficult to apply to real-time protection during everyday exercise for the general public.

[0004] Therefore, in response to the problems mentioned above, this invention proposes a real-time assessment and early warning system for sports injury risk. Summary of the Invention

[0005] To overcome the problems of existing consumer-grade sports monitoring devices having a single risk perception dimension and delayed early warning, while professional analysis systems cannot achieve real-time and convenient daily applications, this invention proposes a real-time sports injury risk assessment and early warning system to achieve proactive and preventative sports injuries.

[0006] The technical solution of the present invention is: a real-time assessment and early warning system for sports injury risk, including a smart sports bracelet and a remote analysis server connected to it. The smart fitness tracker, worn on the user's body, includes: The data acquisition module is used to continuously collect the user's real-time motion physiological data, which includes at least kinematic parameters from the inertial measurement unit, heart rate data from the optical heart rate sensor, and skin conductance activity data from the bioelectrical impedance or skin conductance sensor. The primary processing and communication module is used to perform local preprocessing on the collected real-time motion physiological data and wirelessly transmit the preprocessed data to a remote analysis server. The early warning execution module is used to receive and execute early warning instructions from a remote analysis server, wherein the early warning instructions include at least visual alarm signals corresponding to different risk levels; The remote analysis server includes: The deep analytics engine receives preprocessed data and performs the following analyses: A1, based on heart rate data and its temporal variation characteristics, combined with the user's resting heart rate and the maximum heart rate estimated by age, calculates the cardiac load coefficient; A2, based on the baseline drift characteristics of skin conductance data and the activity intensity in kinematic parameters, estimates the user's fluid balance index through a pre-defined fluid loss correlation model; A3, based on kinematic parameters, uses a built-in motion pattern recognition algorithm to extract stability features reflecting the coordinated activation pattern of muscle groups under the user's current motion posture and abnormal fluctuation features reflecting the intensity of changes in joint motion angles. The risk decision-making module is used for: B1 integrates multimodal data such as cardiac load coefficient, fluid balance index, muscle synergistic activation mode stability characteristics, and joint movement abnormal fluctuation characteristics. B2, based on the fusion results, query the preset multi-level risk thresholds that match the user's basic personal attributes and exercise type; B3. Based on the query results, a current exercise risk level assessment is generated, which includes at least five risk levels from low to high: green (safe), yellow (caution), orange (warning), and red (high risk). The strategy delivery module is used to generate corresponding early warning instructions and supplementary suggestions containing specific intervention measures based on risk level determination, and push them to the early warning execution module of the smart wearable device.

[0007] Preferably, the inertial measurement unit in the data acquisition module includes at least a three-axis accelerometer and a three-axis gyroscope, wherein the kinematic parameters include at least limb acceleration, angular velocity and limb posture angle calculated by a sensor fusion algorithm; It is worth noting that the motion pattern recognition algorithm is based on kinematic parameters. It quantifies the stability characteristics of muscle synergistic activation patterns by analyzing the principal component changes of the covariance matrix of multivariate feature vectors composed of limb acceleration, angular velocity and posture angle within a specific motion cycle. At the same time, it identifies abnormal fluctuation characteristics of joint activity by detecting abrupt changes or high-frequency oscillations that exceed the normal physiological range in the joint angle time series.

[0008] As a preferred option, the construction of the body fluid loss association model specifically includes: establishing the mapping relationship between the low-frequency component change trend of skin electrophysiological activity data and exercise sweating rate, and compensating and correcting the estimated body fluid loss by integrating environmental temperature data and exercise intensity, thereby outputting the body fluid balance state index. It is worth noting that the ambient temperature data is provided by the temperature sensor built into the smart wearable device or by the geographic location and meteorological service data obtained from the connected smart terminal through the communication module.

[0009] Preferably, the deep analysis engine includes a deep learning model based on a temporal neural network. This model takes preprocessed temporal data streams as input, uses cardiac load coefficient, fluid balance index, and muscle co-activation mode stability features as implicit representations in the intermediate layers, and directly outputs a fusion feature vector for risk decision-making. The remote analysis server also includes a user profile database for storing and continuously updating users' basic personal attributes, historical exercise data, personalized risk threshold adjustment parameters, and calibrated deep learning model personalized weights. When querying risk thresholds and calling analysis models, the risk decision-making module prioritizes calling personalized parameters that match the current user profile.

[0010] Preferably, when the warning execution module executes the warning command, for the yellow attention level and higher risk level, the visual alarm signal is accompanied by tactile vibration feedback of different frequencies and patterns, and for the orange warning level and red high-risk level, an additional auditory alarm is triggered. It is worth noting that supplementary recommendations, depending on the risk level and triggering cause, include at least one of the following: recommending replenishment of specific volume of water, recommending reduction of exercise speed or intensity, recommending adjustment of specific limb movement postures, and recommending immediate cessation of exercise and rest.

[0011] Preferably, the remote analysis server also includes a long-term analysis module, which is used to perform statistical analysis on the risk assessment results of a user's single and multiple exercises, generate an exercise safety report reflecting the user's weaknesses (such as specific muscle groups being prone to fatigue or specific joint activity patterns being abnormally frequent), and provide periodic training load adjustment suggestions.

[0012] The beneficial effects of this invention are: 1. This invention uses a smart wearable device to synchronously collect multi-dimensional physiological data such as heart rate, kinematic parameters, and skin electrical activity, and performs real-time fusion processing in a cloud-based deep analysis engine. This transforms the intangible muscle and joint states into the analysis of measurable kinematic parameters (such as muscle synergistic activation stability and abnormal fluctuations in joint activity), thereby enabling real-time perception and assessment of multiple injury risk factors such as cardiac load, dehydration risk, and muscle fatigue.

[0013] 2. This invention sets up complex risk assessment capabilities (such as fatigue recognition and motion analysis based on multi-sensor data fusion) that originally existed only in professional laboratories on everyday smart wearable devices and cloud service platforms. By combining lightweight front-end data collection with powerful cloud-based intelligent analysis, users can obtain near-professional-level injury risk insights and immediate intervention guidance in daily sports scenarios.

[0014] 3. The risk decision-making module of this invention performs risk classification and early warning based on the deeply fused multimodal features and queries personalized thresholds that match the user's personal attributes and exercise type. It can provide differentiated alerts and specific intervention suggestions (such as targeted hydration and posture adjustment) from awareness to high risk. This precise early warning and targeted guidance based on personalized models greatly improves the effectiveness of sports risk management. Attached Figure Description

[0015] Figure 1 The diagram shown is a schematic representation of the system framework of the present invention. Figure 2 The diagram shown is a schematic of the cardiac load coefficient calculation process of the present invention; Figure 3 The diagram shown is a schematic of the fluid balance index estimation process of the present invention. Figure 4 The diagram shown is a schematic of the muscle and joint state feature extraction process of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 some embodiments of the present invention, but 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.

[0017] Please see Figure 1 This invention provides an embodiment of a real-time sports injury risk assessment and early warning system: In this embodiment, the data acquisition module will be described in detail: The module's optical heart rate sensor uses a photoplethysmography (PPG) sensor, which collects pulsation signals from capillaries in the wrist at a specific frequency (50Hz) for subsequent calculation of real-time heart rate and primary indicators of heart rate variability.

[0018] The inertial measurement unit of this module integrates a three-axis MEMS accelerometer and a three-axis MEMS gyroscope. The accelerometer measures the linear acceleration of the limb (unit: g), and the gyroscope measures the angular velocity of the limb (unit: ° / s).

[0019] The module's bioelectrical impedance / electrodermal activity sensor measures changes in electrical impedance between two points on the skin by applying a weak, imperceptible alternating current. Skin electrical activity is closely related to sweat gland activity; its low-frequency, slowly varying component (skin conductance level) effectively reflects sympathetic nerve excitability and is associated with body temperature regulation and sweating, making it a sensitive indicator for indirectly assessing body fluid status.

[0020] This module can also integrate a temperature sensor to monitor wrist skin temperature or ambient temperature, and an integrated barometer to assist in monitoring altitude changes, for use in sports such as mountaineering.

[0021] In this embodiment, the primary processing and communication module will be described in detail: This module runs a lightweight algorithm to perform moving average filtering and outlier removal on the PPG signal, calculate the heart rate value per second, and perform coordinate alignment and zero-bias calibration on the raw data from the inertial measurement unit. It also performs bandpass filtering (e.g., 0.01-0.5Hz) on the electrodermal activity signal to separate the skin conductance level component. The processed data is then timestamped and assembled into data frames (e.g., one frame per second, containing real-time heart rate, triaxial acceleration, triaxial angular velocity, skin conductance level, etc.), which are transmitted via Bluetooth to a paired smartphone. The smartphone application then forwards the data to a remote server via the mobile internet.

[0022] In this embodiment, the early warning execution module is described in detail: This module includes a colored LED indicator (or OLED screen), a linear vibration motor, and a miniature speaker. It receives instructions from the server and interprets the risk level (e.g., "Level": "Orange") and motion code. For example, a yellow level triggers a soft flashing yellow light and two short vibrations; an orange level triggers a rapid flashing orange light and strong intermittent vibrations; and a red level adds a rapid "beep" alarm sound to the above. Simultaneously, a prominent warning pops up at the top of the connected phone screen, displaying specific suggestions such as "Warning: A significant decrease in motion stability has been detected. It is recommended to immediately reduce pace and pay attention to landing posture."

[0023] In this embodiment, the remote analysis server will be described in detail: (1) The server's deep analytics engine concurrently processes data from a massive number of users, and performs the following analysis independently on each user's data, specifically: Please see Figure 2 For the calculation of cardiac load factor: S1, first input the preprocessed heart rate time series data.

[0024] S2, then estimate the maximum heart rate based on the user's registration information (age). ≈ 220 - age). Calculate the current heart rate reserve percentage (%HRR = [( - ) / ( - ()] × 100%, where For the user's daily resting heart rate, The user's current heart rate is used. The recent (e.g., past 2 minutes) heart rate rise slope (ΔHR / Δt) is also calculated. The cardiac load factor (CLC) is defined as: CLC = , where α and β are weighting coefficients, and α + β = 1. %HRR reflects absolute load, and the rising slope reflects the rate of load increase. For example, a state that rapidly reaches a high %HRR is more risky than the same level reached slowly.

[0025] S3 outputs a continuous value between 0 and 1 (or 0 and 100). The higher the value, the greater the risk of instantaneous pressure and cumulative load on the heart.

[0026] Please see Figure 3 For the estimation of the fluid balance index: S1: First, input the skin conductance level time series data, the motion intensity (such as acceleration vector amplitude) obtained from the inertial measurement unit, and the ambient temperature.

[0027] S2, the pre-defined fluid loss correlation model, is a physiological model calibrated with experimental data. The core of this model is a strong positive correlation between the upward trend of skin conductivity and sweating rate. The model first estimates the basal sweating rate based on the rate of increase of skin conductivity after the start of exercise and its current value. Then, it incorporates exercise intensity compensation (high-acceleration activities (such as sprinting) induce faster sweating). Next, it incorporates ambient temperature compensation (in high-temperature environments, the actual amount of sweat evaporation corresponding to the same skin conductivity level may be greater, leading to faster dehydration). The model integrates these factors to estimate the "effective fluid equivalent" since the start of exercise. The Fluid Balance Index (FBI) is calculated as: FBI = Initial estimated total water reserves (based on body weight) - Cumulative effective fluid equivalent. This index is expressed in milliliters (ml) or as a percentage.

[0028] S3 outputs a dynamically decreasing value, indicating an increased risk of dehydration when it falls below a personalized threshold (such as 1% of body weight loss). This model can detect early signs of dehydration earlier than simply experiencing thirst.

[0029] Please see Figure 4For muscle and joint state feature extraction: S1 takes inertial measurement unit data processed by sensor fusion algorithms (such as complementary filtering or Kalman filtering) as input and outputs more accurate limb attitude angles (pitch, roll, yaw) and calibrated acceleration and angular velocity.

[0030] S2 extracts stability features from muscle synergistic activation patterns. The system has a built-in standard pattern library for various common sports (such as running, cycling, and aerobics). For the current sport, the algorithm first performs pattern recognition (e.g., recognizing the user as "running"), and then extracts a multivariate feature vector consisting of the posture angles, angular velocities, and accelerations in specific directions at key points within a complete gait cycle. Using principal component analysis or similar methods, the projection of the current cycle feature vector onto the feature space spanned by the user's historical normal data or a standard model is calculated. The stability feature is quantified as the Mahalanobis distance or cosine similarity between the current projection and the "ideal" or "average" projection. An abnormally high stability feature value indicates a decline in neuromuscular control efficiency and a "drift" or "mismatch" in the muscle group synergistic working pattern. This is a sensitive indicator of over-fatigue or early unconscious compensation, often appearing before obvious pain or strength loss.

[0031] Simultaneously, abnormal fluctuation features of joint activity are extracted. First, for the identified major joints (such as the knee joint during running and the elbow joint during weightlifting), the algorithm continuously tracks the calculated joint angles. By analyzing the time series of joint angles, two types of anomalies are detected: First, drastic changes in angles exceeding the normal physiological range within a very short period (possibly indicating emergency stop, twisting, or loss of control); second, an abnormally increased number of small-amplitude high-frequency jitters in the angle signal (possibly corresponding to tremors or instability caused by muscle fatigue). Abnormal fluctuation features are quantified as the number of abnormal events detected per unit time or the accumulated energy.

[0032] S3 outputs stability characteristics (continuous values, the lower the value, the more stable) and abnormal fluctuation characteristics (discrete event count or intensity index).

[0033] (2) The risk decision-making module of this server is described in detail: This module receives cardiac load coefficients, fluid balance indexes, stability characteristics, and abnormal fluctuation characteristics from the deep analysis engine, and then performs data fusion using weighted fusion or a rule engine based on fuzzy logic, for example, to build a risk score (RS) model: ; Norm() is a function that normalizes each indicator to a risk score of [0,1]. - The weights are dynamic and can be adjusted based on the type of exercise (e.g., more emphasis is placed on long-distance running). and Strength training focuses more on and ).

[0034] This module also features a user profile database, which includes age, gender, weight, exercise habits, injury history, and personal baselines derived through long-term learning (such as the individual's normal stability characteristic range and typical sweating rate). Risk thresholds are not fixed but are tied to the user profile and the current type of exercise. For example, an experienced marathon runner's "orange" cardiac load coefficient threshold may be higher than that of a novice runner.

[0035] This module determines the real-time risk level based on the fused RS value or personalized rules (such as "immediately upgrade to orange if abnormal fluctuation characteristics exceed the threshold twice"). Specifically: Green (safe), all indicators are within the range of personal safety.

[0036] Yellow (Note): 1-2 indicators are close to or slightly exceed the threshold, and the risk is controllable.

[0037] Orange (Warning): Core indicators (such as cardiac load coefficient and stability characteristics) significantly exceed the threshold, or a single indicator (such as a sharp increase in abnormal fluctuation characteristics) shows an increased acute risk and a significantly increased likelihood of injury.

[0038] Red (high risk): Multiple indicators are severely exceeded, or a pattern that is very likely to cause immediate injury is detected (such as a sudden change in knee joint angle suspected to be a sprain). You need to stop immediately.

[0039] (3) The policy distribution module of the server generates precise warning instructions and text suggestions based on the judgment level and specific triggering indicators. For example, the yellow warning triggered by the fluid balance index suggests: "You have sweated a lot. It is recommended to replenish 150ml of water within 5 minutes"; the orange warning triggered by the stability feature suggests: "A decrease in running posture stability has been detected (possibly due to fatigue). It is recommended to immediately reduce the pace and focus on core tightening and gentle landing."

[0040] (4) The long-term analysis module of the server generates a detailed report for the user after the exercise ends, pointing out the period with the highest risk and the main risk factors in this exercise, and giving long-term suggestions, such as "Your right knee stability showed a significant downward trend in the later stage of your run this month, and it is recommended to strengthen the right hip muscle group strength training."

[0041] This invention provides an embodiment 1: In this embodiment, the invention is applied to a 30-year-old amateur runner (weighing 70kg), while wearing a traditional single heart rate bracelet on the other hand. The test subject is running 10 kilometers in the morning at a temperature of 25°C. The individual dehydration warning threshold is set to a fluid loss of 0.5L (approximately 0.7% of body weight).

[0042] Working process of this invention: S1, 20 minutes after the start of the run, the system estimated that the body fluid loss was close to 0.4L based on the rising trend of skin conductivity and exercise intensity. The body fluid balance index dropped, triggering a yellow warning: "Pay attention to replenishing water, about 400ml of water has been lost."

[0043] S2, after running for 35 minutes, the heart rate reached 92% of the maximum heart rate and the rate of increase was fast, indicating a high cardiac load coefficient; at the same time, the data analysis of the inertial measurement unit showed that the trunk rotation stability (stability characteristics) in the gait cycle began to decline. The system combined the two information and judged it as an orange warning: "High cardiac load and decreased postural stability, it is recommended to reduce pace and adjust breathing."

[0044] S3: Runners follow the advice to slow down. The system continues to monitor the runners, and as their heart rate decreases and their posture stability improves, the risk level returns to yellow.

[0045] Traditional single heart rate wristbands only alarm when the heart rate exceeds a preset fixed threshold (such as 180 beats / minute). During the entire process, although the runner's heart rate is high, it does not exceed its fixed threshold, so there is no warning. The runner may continue to persist without replenishing water and in a state of accumulated fatigue, increasing the risk of dehydration, heatstroke, or patellar syndrome or stress fracture caused by fatigue-induced movement deformities.

[0046] This invention provides an embodiment 2: In this embodiment, the invention is applied to a 24-year-old fitness novice, and the test subject performs bench press training to verify the invention.

[0047] Working process of this invention: In the second half of the third training session, the system's data analysis using the inertial measurement unit revealed that when pushing the barbell, the synchronicity of the angular velocity of the elbow joints on both sides deteriorated (stability characteristics decreased), and the right shoulder joint exhibited slight high-frequency shaking (abnormal fluctuation characteristic event). This indicates that the pectoralis major and anterior deltoid muscles may be approaching exhaustion, and asymmetric compensation and core instability are beginning to occur.

[0048] S2, the system immediately triggered an orange warning: "Asymmetrical movements and shaking have been detected. It is recommended to immediately terminate this set, rest for a longer period of time, or reduce the weight."

[0049] As can be seen from this embodiment, the analysis of the present invention is online and timely. Compared with professional laboratory surface electromyography systems, although they can accurately detect the imbalance of electromyographic activity of the left and right pectoralis major muscles and shoulder shaking, their analysis is offline. Data collection requires several hours or even days to process and analyze, and cannot provide protection on the spot during training. Novices may complete dangerous movements in a state of exhaustion, resulting in rotator cuff injury or accidental crushing injury.

[0050] This invention provides an embodiment 3: In this embodiment, the invention is applied to a 26-year-old athlete, who undergoes intermittent training at a high temperature of 35°C to verify the invention.

[0051] Working process of this invention: S1, In this embodiment, the ambient temperature is introduced into the body fluid loss correlation model of the present invention as a key input. Only 15 minutes after the start of exercise, due to the high temperature, the sweating rate surges and the body fluid balance index estimated by the system drops rapidly, triggering an orange warning: "Water loss is extremely fast in high temperature environment. Please stop immediately and replenish with 300ml of electrolyte beverage."

[0052] S2 was also monitored and found that due to the increase in core temperature and early dehydration, the heart rate recovery rate was significantly slower during the interval, and the cardiac load coefficient remained at a high level.

[0053] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A real-time sports injury risk assessment and early warning system, characterized in that, This includes smart wearable devices and remote analysis servers that communicate with them; The smart wearable device, worn on the user's limbs, includes: The data acquisition module is used to continuously collect the user's real-time motion physiological data, which includes at least kinematic parameters from the inertial measurement unit, heart rate data from the optical heart rate sensor, and skin conductance activity data from the bioelectrical impedance or skin conductance sensor. The primary processing and communication module is used to perform local preprocessing on the collected real-time motion physiological data and wirelessly transmit the preprocessed data to a remote analysis server. The early warning execution module is used to receive and execute early warning instructions from a remote analysis server, wherein the early warning instructions include at least visual alarm signals corresponding to different risk levels; The remote analysis server includes: The deep analytics engine receives preprocessed data and performs the following analyses: A1, based on heart rate data and its temporal variation characteristics, combined with the user's resting heart rate and the maximum heart rate estimated by age, calculates the cardiac load coefficient; A2, based on the baseline drift characteristics of skin conductance data and the activity intensity in kinematic parameters, estimates the user's fluid balance index through a pre-defined fluid loss correlation model; A3, based on kinematic parameters, uses a built-in motion pattern recognition algorithm to extract stability features reflecting the coordinated activation pattern of muscle groups under the user's current motion posture and abnormal fluctuation features reflecting the intensity of changes in joint motion angles. The risk decision-making module is used for: B1 integrates multimodal data such as cardiac load coefficient, fluid balance index, muscle synergistic activation mode stability characteristics, and joint movement abnormal fluctuation characteristics. B2, based on the fusion results, query the preset multi-level risk thresholds that match the user's basic personal attributes and sports type; B3. Based on the query results, a current exercise risk level assessment is generated, which includes at least five risk levels, from low to high: green (safe), yellow (caution), orange (warning), and red (high risk). The strategy delivery module is used to generate corresponding early warning instructions and supplementary suggestions containing specific intervention measures based on risk level determination, and push them to the early warning execution module of the smart wearable device.

2. The real-time sports injury risk assessment and early warning system according to claim 1, characterized in that: The inertial measurement unit in the data acquisition module includes at least a three-axis accelerometer and a three-axis gyroscope, wherein the kinematic parameters include at least limb acceleration, angular velocity and limb posture angle calculated by a sensor fusion algorithm.

3. The real-time sports injury risk assessment and early warning system according to claim 1, characterized in that, The construction of the fluid loss correlation model specifically includes: establishing a mapping relationship between the low-frequency component change trend of skin electrical activity data and exercise sweating rate, and compensating and correcting the estimated fluid loss by integrating environmental temperature data and exercise intensity, thereby outputting a fluid balance state index.

4. The real-time sports injury risk assessment and early warning system according to claim 1, characterized in that, The extraction of the stability features of the muscle synergistic activation mode is specifically as follows: within a preset standard movement cycle, the principal component changes of the covariance matrix of a specific feature vector composed of limb acceleration and angular velocity are analyzed, and the degree of deviation from the typical pattern is quantified as a stability feature.

5. The real-time sports injury risk assessment and early warning system according to claim 1, characterized in that: The deep analysis engine includes a deep learning model based on a temporal neural network. This model takes preprocessed temporal data streams as input, uses cardiac load coefficient, fluid balance index and muscle co-activation mode stability features as implicit representations in the intermediate layer, and directly outputs a fusion feature vector for risk decision-making.

6. The real-time sports injury risk assessment and early warning system according to claim 1, characterized in that: The remote analysis server also includes a user profile database, which stores and continuously updates users' basic personal attributes, historical exercise data, and personalized risk threshold adjustment parameters; when querying risk thresholds, the risk decision module prioritizes calling personalized thresholds that match the current user profile.

7. The real-time sports injury risk assessment and early warning system according to claim 1, characterized in that: When the warning execution module executes the warning command, for the yellow alert level and higher risk level, the visual alarm signal is accompanied by tactile vibration feedback of different frequencies and patterns. For the orange warning level and the red high-risk level, an additional auditory alarm is triggered.

8. The real-time sports injury risk assessment and early warning system according to claim 1, characterized in that, The supplementary suggestions generated by the strategy delivery module, based on different risk levels and triggering reasons, include at least one of the following: suggesting replenishing a specific volume of water, suggesting reducing the speed or amplitude of exercise, suggesting adjusting specific limb movements and postures, and suggesting immediately stopping exercise and resting.

9. The real-time sports injury risk assessment and early warning system according to claim 1, characterized in that: The remote analysis server also includes a long-term analysis module, which is used to statistically analyze the risk assessment results of a user's single and multiple exercises, generate an exercise safety report reflecting the user's weaknesses, and provide periodic training load adjustment suggestions.

10. The real-time sports injury risk assessment and early warning system according to claim 1, characterized in that: The smart wearable device is a smart fitness tracker.

Images