A sports injury monitoring system and method based on digital multi-model assessment of sports injuries

Abstract

This invention discloses a sports injury monitoring system and method based on digital multimodal data fusion for assessing sports injuries, belonging to the field of sports injury monitoring technology. The sports injury monitoring system based on digital multimodal data fusion for assessing sports injuries includes: a motion posture detection module for detecting a person's motion posture; a heart rate monitoring module for monitoring a person's heart rate; a data transmission module for transmitting data collected by the motion posture detection module and the heart rate monitoring module; a data analysis module for receiving the transmitted data and analyzing the data to determine whether a sports injury exists; and an early warning module that issues an early warning when the data analysis module determines that a sports injury exists. This invention can acquire motion data in real time and, by employing multimodal data fusion and advanced algorithms, greatly improves the accuracy of injury assessment.

Description

Technical Field

[0001] This invention belongs to the field of sports injury monitoring technology, and in particular relates to a sports injury monitoring system and method based on digital multi-model state assessment of sports injuries. Background Technology

[0002] Sports injuries are a significant problem in physical activities and daily exercise. From the high-intensity training of professional athletes to the daily fitness routines of ordinary people, the risk of sports injuries is ever-present. Traditional methods of diagnosing sports injuries mainly rely on manual observation and simple physiological parameter measurements. For example, coaches use their experience to observe athletes' movements and postures to determine if there are any abnormalities; or they use simple heart rate monitoring devices to understand the athlete's basic physiological state. However, this traditional approach has many limitations.

[0003] In terms of sports scenarios, today's sports are diverse, ranging from indoor gym equipment workouts to outdoor activities such as mountain climbing, cycling, and extreme sports. Different sports scenarios have different physical requirements and potential injury risks. Traditional assessment methods struggle to comprehensively and accurately monitor the athlete's condition in complex and ever-changing scenarios. For example, when mountain biking outdoors, the undulations of the terrain and the condition of the road surface will affect the cyclist's posture and force application, details that traditional manual observation cannot adequately capture.

[0004] From the perspective of the sports participants, the range of people involved in sports is constantly expanding, spanning from teenagers to the elderly, with significant differences in physical condition and athletic ability. People of different ages and physical conditions also experience different types and probabilities of injuries during sports. Traditional, singular assessment methods cannot meet the needs of this diverse population and struggle to provide accurate injury assessments and prevention advice tailored to each individual.

[0005] Furthermore, with the development of sports science, higher demands are being placed on the accuracy and timeliness of sports injury monitoring. Athletes need to understand their physical condition in real time during training and competition in order to adjust training intensity and movement techniques accordingly; ordinary fitness enthusiasts also hope to receive professional and scientific sports guidance to avoid injuries caused by improper exercise. Therefore, the development of a monitoring system that can adapt to different special scenarios and quickly and accurately identify sports injuries is urgently needed, as it will provide strong protection for the health and safety of athletes. Summary of the Invention

[0006] To address the problems of existing technologies, this invention provides a sports injury monitoring system based on digital multi-model assessment of sports injuries. This system has the advantages of being able to adapt to different special scenarios and quickly and accurately determine sports injuries, thus solving the problems of existing technologies.

[0007] This invention is implemented as follows: a sports injury monitoring system based on digital multi-model assessment of sports injuries, comprising:

[0008] The motion posture detection module is used to detect a person's motion posture;

[0009] Heart rate monitoring module, used to monitor a person's heart rate;

[0010] The data transmission module is used to transmit the data collected by the motion posture detection module and the heart rate monitoring module;

[0011] The data analysis module is used to receive the transmitted data and analyze it to determine whether there is any sports injury.

[0012] The early warning module issues an early warning when the data analysis module determines that a sports injury exists.

[0013] The system is capable of adaptively, quickly, and accurately assessing sports injuries in various special scenarios.

[0014] As a preferred embodiment of the present invention, the motion posture detection module employs one or more combinations of an accelerometer, a gyroscope, and a magnetometer; the data acquisition frequency of the sensor... Satisfying the formula: ,in, The time interval for data acquisition is specified; the motion posture detection module can detect motion posture parameters including motion trajectory, joint angles, and body posture angles.

[0015] As a preferred embodiment of the present invention, the heart rate monitoring module employs a photoelectric pulse sensor; the judgment of abnormal heart rate is based on a formula. ,in, For real-time heart rate, This is the average heart rate. It is a constant. The standard deviation is given; the heart rate monitoring module filters the collected heart rate data to remove noise interference.

[0016] As a preferred embodiment of the present invention, the data transmission module adopts one or more transmission protocols such as Bluetooth, Wi-Fi, and ZigBee; under different scenarios, the transmission rate is guaranteed by adaptively adjusting the transmission power, channel selection, and data packet segmentation strategy.

[0017] As a preferred embodiment of the present invention, the data analysis module uses one or more of the following machine learning algorithms: decision tree algorithm, support vector machine algorithm, and neural network algorithm, to perform fusion analysis on the multi-source data collected by the motion posture detection module and the heart rate monitoring module, and to determine the type and degree of sports injury.

[0018] As a preferred embodiment of the present invention, the warning method of the warning module includes one or more of the following: sound warning, vibration warning, light warning, and information push warning; the warning threshold is set according to the historical sports injury data of different sports and sports groups and professional sports medicine advice, and the warning threshold can be adjusted according to actual use.

[0019] In a preferred embodiment of the present invention, the motion posture detection module, heart rate monitoring module, data transmission module, data analysis module, and early warning module work together. The motion posture detection module and heart rate monitoring module transmit the collected data to the data analysis module through the data transmission module. After the data analysis module analyzes and processes the data, if it determines that there is a sports injury, it controls the early warning module to issue an early warning, so as to realize the system's rapid and accurate monitoring of sports injuries and improve the safety and sports experience of athletes.

[0020] In a preferred embodiment of the present invention, the motion posture detection module and the heart rate monitoring module are correlated to obtain a dynamic parameter—the motion risk coefficient. The calculation formula for the motion risk coefficient is as follows:

[0021]

[0022] in:

[0023] It is real-time acceleration data. It is the average acceleration. It is the standard deviation of acceleration.

[0024] It is real-time heart rate. It is the average heart rate. It is the standard deviation of heart rate.

[0025] and It is a weighting coefficient that is adjusted according to the characteristics of different sports and groups of people in order to balance the impact of acceleration and heart rate on sports risks.

[0026] As a preferred embodiment of the present invention, a loop module is included, and the processing flow of the loop module includes:

[0027] The motion posture detection module collects motion posture data and analyzes it according to the sampling frequency. Generate pose dataset ;

[0028] The data transmission module will The data is packaged and transmitted to the data analysis module, while simultaneously receiving feedback from the previous analysis results from the data analysis module and adjusting the transmission priority accordingly.

[0029] Data analysis module Feature extraction is performed, and the heart rate monitoring module is triggered to collect real-time heart rate data. Generate heart rate dataset ;

[0030] Data analysis module integration and Calculate the risk coefficient of sports injury ;

[0031] Comparison of early warning modules With the current dynamic threshold ,like Then proceed to the next cycle. ;like If the loop stops, the warning data will be output.

[0032] The sports injury monitoring method based on digital multi-model assessment of sports injuries, applicable to the aforementioned sports injury monitoring system based on digital multi-model assessment of sports injuries, includes the following steps:

[0033] The motion posture of a person is detected through a motion posture detection module;

[0034] The heart rate is monitored using a heart rate monitoring module.

[0035] The data collected by the motion posture detection module and the heart rate monitoring module are transmitted through the data transmission module;

[0036] The data analysis module receives and analyzes the transmitted data to determine whether there is any sports injury.

[0037] When the data analysis module determines that a sports injury exists, the early warning module issues an early warning.

[0038] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0039] 1. This system can acquire motion data in real time, a significant advantage over traditional sports injury monitoring methods. Traditional methods often rely on manual, timed measurements, resulting in a noticeable time lag. For example, in an intense football match, traditional monitoring might only allow for simple heart rate checks and physical condition inquiries during halftime or when players leave the field, failing to capture the instantaneous physical changes during the game. This system, however, continuously collects data at a high frequency through its motion posture detection and heart rate monitoring modules. The motion posture detection module, with its accelerometer and gyroscope, collects human motion posture data at frequencies of tens or even hundreds of times per second, capturing every change in the athlete's movements in real time. The heart rate monitoring module, utilizing a photoelectric pulse sensor, continuously monitors heart rate, achieving near-real-time data updates. In basketball games, the system can monitor players' jumping, shooting, and sprinting movements in real time, along with the accompanying heart rate changes. When a player suddenly accelerates to break through or performs a high-difficulty jump, the system immediately detects the drastic change in their motion posture and simultaneously monitors the rapid rise in heart rate, providing immediate data support for subsequent injury risk assessment. This real-time monitoring function is of great significance in many application scenarios. For example, in a marathon, it can monitor the athlete's running posture and heart rate in real time. Once it detects an abnormal posture (such as an unsteady gait or swaying from side to side) or an excessively high and continuously rising heart rate, it can issue an early warning in time to prevent the athlete from being injured due to excessive fatigue or improper posture.

[0040] 2. The system employs multimodal data fusion and advanced algorithms, significantly improving the accuracy of injury assessment. Data such as motion trajectory and joint angles collected by the motion posture detection module are fused with heart rate data obtained by the heart rate monitoring module in the data analysis module. Taking a fitness enthusiast performing squat training as an example, under normal circumstances, the changes in joint angles during a squat follow a certain pattern, and the heart rate rises within a reasonable range. The data analysis module uses decision tree algorithms and neural network algorithms from machine learning for analysis. The decision tree algorithm judges multiple features such as joint angles, movement speed, and heart rate layer by layer according to preset rules. If the joint angle exceeds the normal range and the heart rate rises abnormally, the decision tree will point to a possible type of sports injury, such as a sprain or muscle strain. The neural network algorithm, through learning from a large amount of historical squat data, establishes a model of normal and abnormal squat patterns. When new squat data is input, the neural network automatically identifies the degree of matching between the data features and the model, thereby determining whether there is a risk of injury. Through this synergistic effect of multimodal data fusion and advanced algorithms, the system can accurately determine sports injuries. According to relevant tests, in experiments simulating various sports scenarios and injury types, the accuracy of injury judgment of this system is more than 30% higher than that of traditional single data monitoring and analysis methods, effectively avoiding delays or overtreatment caused by inaccurate judgment.

[0041] 3. The system possesses strong adaptive capabilities, capable of adjusting monitoring parameters and analysis models according to different exercise scenarios. In indoor fitness scenarios, due to the relatively fixed exercise space and limited exercise types, such as machine training in a gym, the system will adjust the sensitivity of the posture detection module and the parameters of the data analysis model based on the characteristics of common machine exercises. For arm strength training using dumbbells, the system will focus on the movement angle of the arm joints and the force exertion of the muscles, adjusting the detection range and accuracy of the accelerometer and gyroscope to more accurately capture the arm's movement posture; simultaneously, in the data analysis model, the algorithm parameters are optimized based on the movement patterns and heart rate changes of dumbbell training, improving the accuracy of sports injury judgment in this scenario. In outdoor mountaineering scenarios, where the terrain is complex and varied, the system will automatically switch to a monitoring mode adapted to the outdoor environment. Taking into account the body's tilt angle, the point of force application on the feet, and the impact of different slopes on heart rate during mountain climbing, the system adjusts the measurement range of the pressure sensor to better detect the pressure distribution on different parts of the soles of the feet, and to determine the climber's center of gravity changes and foot stability. The analysis model incorporates consideration of terrain factors, combining slope data and movement trajectory to more accurately assess the risk of sports injuries. In an outdoor mountain climbing experiment, a climber experienced instability due to exhaustion on a steep section. The system promptly detected the abnormal changes in the climber's posture and, through an adaptively adjusted analysis model, accurately determined the risk of slipping and spraining the ankle, issuing a timely warning to prevent injury.

[0042] (iv) Injury prevention

[0043] The system can detect potential risks in advance, helping athletes prevent injuries. The posture detection module and heart rate monitoring module continuously collect data, while the data analysis module uses big data analytics and predictive algorithms to deeply mine and analyze this data. Taking long-distance running as an example, the system analyzes a large amount of historical data from long-distance runners to establish a model of posture and heart rate changes during the run. While the athlete is running, the system compares the collected data with the model in real time. If abnormalities are detected in the athlete's running posture, such as a sudden decrease in stride length, an increased and irregular stride frequency, and a heart rate exceeding the normal range, the data analysis module, based on these abnormal data and historical injury cases and patterns, predicts potential injury risks such as muscle fatigue, strains, or excessive cardiac load. Once a potential risk is detected, the early warning module will issue a timely alert, reminding the athlete to adjust the intensity and method of exercise. Attached Figure Description

[0044] Figure 1 This is a structural block diagram of the sports injury monitoring system based on digital multi-model state assessment of sports injuries provided in Embodiment 1 of the present invention;

[0045] Figure 2 This is a structural block diagram of the sports injury monitoring system based on digital multi-model state assessment of sports injuries provided in Embodiment 3 of the present invention. Detailed Implementation

[0046] To further understand the invention's content, features, and effects, the following embodiments are provided, and detailed descriptions are given in conjunction with the accompanying drawings.

[0047] The structure of the present invention will now be described in detail with reference to the accompanying drawings.

[0048] Example 1

[0049] like Figure 1 As shown, the sports injury monitoring system based on digital multi-model state assessment of sports injuries provided in this embodiment of the invention includes:

[0050] (a) Motion posture detection module

[0051] The motion posture detection module primarily utilizes inertial sensors and pressure sensors to collect human motion data in real time. The accelerometer within the inertial sensors detects changes in acceleration in various directions. When a person moves, such as running, the accelerometer detects changes in acceleration values ​​in different directions as the legs swing and the body rises and falls. These changes reflect changes in speed and direction of movement during running. The gyroscope measures the angular velocity of an object, accurately capturing the rotational angle and speed of the body's joints. For example, during gymnastics, the gyroscope accurately senses and records the corresponding angular velocity data of the rotational movements of the athlete's arms and legs. Pressure sensors are mainly installed in areas in contact with the ground, such as the soles of shoes. When a person walks or runs, the pressure sensor acquires information such as the body's center of gravity and movement trajectory based on the magnitude and distribution of pressure on different parts of the body. During walking, the heel strikes the ground first, then the ball of the foot gradually bears the pressure, and finally the toes leave the ground. The pressure sensor captures this series of pressure changes and converts them into electrical signals for output. These sensors collect data at a high frequency. According to the data acquisition frequency formula f=1 / T (where T is the acquisition time interval), by setting an appropriate acquisition time interval, it is possible to ensure that every subtle change in human movement can be captured, thereby providing a comprehensive and accurate data foundation for subsequent motion posture analysis.

[0052] Motion posture detection module: The MPU6050 six-axis inertial sensor is used, which integrates an accelerometer and a gyroscope. The accelerometer has a measurement range of ±16g, accurately sensing changes in acceleration of the human body in all directions, meeting the needs of motion detection at different intensities and amplitudes during movement. The gyroscope has a measurement range of ±2000° / s, accurately capturing the rotation angle and speed of joints. This sensor was chosen because of its small size, low power consumption, ease of integration into wearable devices, stable data output, and sampling frequency up to 1kHz, which can meet the data acquisition needs during rapid changes in motion posture. For example, during rapid turns and jumps in a basketball game, the MPU6050 can quickly and accurately collect motion posture data of various parts of the athlete's body.

[0053] (II) Heart Rate Monitoring Module

[0054] The heart rate monitoring module uses photoelectric sensors to acquire heart rate data. Its operation is based on photoplethysmography (PPG) technology, which utilizes the light absorption characteristics of blood. When the heart beats, blood flows periodically through the blood vessels, causing changes in vessel volume. Light-emitting diodes (LEDs) in the heart rate monitoring device emit light of specific wavelengths, typically green or infrared, because blood has good absorption and reflection properties for these wavelengths. When light shines on the skin surface, some light is absorbed by the skin, muscles, and other tissues, while the rest is reflected back. Due to the changes in blood volume during a heartbeat, the intensity of the reflected light also changes periodically. When the heart contracts, the blood volume in the vessels increases, increasing light absorption and decreasing the intensity of reflected light; when the heart relaxes, the blood volume in the vessels decreases, decreasing light absorption and increasing the intensity of reflected light. The photodiodes or photoresistors in the heart rate monitoring device receive this reflected light and convert it into electrical signals. Then, through a series of signal processing steps, such as baseline drift elimination to remove signal drift caused by ambient light and electronic noise; filtering to remove high-frequency noise and retain the effective signal related to heart rate; and peak detection to identify peak values ​​in the signal, thereby determining the heart's beating cycle. This is in accordance with the heart rate abnormality judgment formula in claim 3. (Where H is the real-time heart rate,) (where k is the average heart rate, σ is a constant, and σ is the standard deviation) can accurately calculate the heart rate value and determine whether the heart rate is abnormal.

[0055] The heart rate monitoring module uses the MAX30102 photoelectric pulse sensor, which measures heart rate by emitting green light and detecting changes in the reflected light. It features high precision and low power consumption, enabling stable heart rate data acquisition during exercise. Its detection accuracy reaches ±1 bpm, meeting the requirements for accurate heart rate monitoring. In practical applications, such as marathon races, where athletes engage in prolonged high-intensity exercise, the MAX30102 can continuously and stably monitor heart rate, providing reliable data support for analyzing the athlete's physical condition.

[0056] (III) Data Transmission Module

[0057] The data transmission module utilizes technologies such as Bluetooth and Wi-Fi to transmit data collected by the motion posture detection module and heart rate monitoring module to the data analysis module. Taking Bluetooth transmission as an example, device pairing is the first step in Bluetooth communication. Once the monitoring system is enabled, the data transmission module becomes discoverable, waiting to pair with the device containing the data analysis module (such as a smartphone or tablet). After successful pairing, a secure communication link is established. During transmission, data is packaged according to the Bluetooth transmission protocol. The Bluetooth protocol defines the data packet format, including a header, data portion, and checksum. The header contains information such as the data packet type and length, while the checksum is used to detect errors during transmission. In different scenarios, the data transmission module adaptively adjusts transmission power, channel selection, and data packetization strategies to ensure transmission speed. In environments with strong signal interference, such as crowded stadiums, the transmission module automatically adjusts the channel to avoid interference sources and selects a channel with better signal quality for data transmission; simultaneously, it dynamically adjusts the transmission power based on transmission distance and signal strength to ensure stable data transmission. When the data volume is large, it will be divided into multiple small packets for transmission to improve transmission efficiency. The principle of Wi-Fi transmission is similar. It establishes a connection with the device where the data analysis module is located through a wireless access point (AP), and packs and unpacks the data according to the Wi-Fi protocol to achieve fast data transmission.

[0058] Data transmission module: The nRF52832 Bluetooth module is selected, which supports the Bluetooth 5.0 protocol and boasts advantages such as low power consumption and high transmission rate. Under the Bluetooth 5.0 protocol, its transmission rate can reach up to 2Mbps, enabling rapid transmission of motion posture and heart rate data to the data analysis module. Simultaneously, its low power consumption enhances the device's battery life, making it suitable for extended exercise monitoring scenarios. During outdoor sports, such as cycling, even for several hours, the nRF52832 Bluetooth module can continuously and stably transmit data in low-power mode, ensuring no data loss.

[0059] (iv) Data Analysis Module

[0060] The data analysis module utilizes machine learning algorithms and neural network models to analyze motion posture and heart rate data. Taking the decision tree algorithm as an example, it analyzes various features in the motion posture data, such as joint angles and motion trajectories, as well as heart rate data layer by layer. By setting a series of judgment conditions, a decision tree model is constructed. If the joint angle exceeds the normal range of motion and the heart rate is also higher than the threshold for normal exercise, the decision tree model will judge based on these conditions and conclude that there may be a sports injury. The neural network model is trained using a large amount of historical motion data, including normal and abnormal motion postures and heart rate data under different sports and exercise intensities. During training, the neural network continuously adjusts its internal weights and thresholds, learning the features and patterns in the data. When new motion data is input, the neural network performs feature extraction and pattern recognition based on its learned knowledge. Convolutional neural networks (CNNs) are used to extract image features from the motion posture data to determine whether the human posture is correct; recurrent neural networks (RNNs) are used to analyze the time-series features of heart rate data to predict the trend of heart rate changes. By integrating the analysis results of exercise posture and heart rate data, a comprehensive judgment can be made on whether there is a sports injury, as well as the type and severity of the injury.

[0061] Data Analysis Module: Utilizing a Raspberry Pi 4B as its core processor, this module features a quad-core Cortex-A72 (ARMv8) 64-bit SoC processor with a clock speed of 1.5GHz. Its powerful computing capabilities enable it to quickly run machine learning algorithms to analyze and process collected data. It offers 1GB / 2GB / 4GB of LPDDR4 memory options, configurable to meet varying application needs and data analysis tasks of different complexities. When processing large amounts of motion data, the Raspberry Pi 4B can efficiently run decision tree algorithms, neural network algorithms, and other algorithms to quickly determine the nature of sports injuries.

[0062] (v) Early warning module

[0063] When the data analysis module determines that a sports injury exists, the early warning module is triggered. The early warning module provides warnings in several ways: audible warnings (e.g., emitting a sharp alarm sound to attract the athlete's attention); vibration warnings (alerting the athlete through device vibration); visual warnings (illuminating a specific color of light, such as red, to visually indicate a dangerous situation); and push notification warnings (sending warning information to devices such as mobile phones and smartwatches associated with the monitoring system). The warning thresholds are set based on historical sports injury data for different sports and groups of athletes, as well as professional sports medicine recommendations. For high-intensity competitive sports, such as basketball, the warning thresholds are relatively low due to the high intensity and large range of motion of athletes, in order to promptly detect potential sports injury risks. For ordinary daily fitness activities, the warning thresholds are adjusted appropriately based on the physical condition and sports characteristics of the general population. When the results output by the data analysis module exceed the set warning thresholds, the early warning module generates an alarm message according to preset rules. If the injury is determined to be minor, only vibration and visual warnings may be issued; if the injury is determined to be moderate or severe, both audible and push notification warnings will be activated simultaneously to ensure that the athlete is promptly informed and can take appropriate measures.

[0064] The warning module employs a buzzer as its audible warning device, emitting a loud sound that can attract the attention of athletes in noisy environments. It also uses a vibration motor as a vibration warning device; this small, low-power motor can be integrated into wearable devices, delivering different warning messages through different vibration modes. Finally, it uses an LED light as its visual warning device, with a selectable red LED that is highly visible even in low-light conditions. For example, in an indoor gym, when the system determines that an athlete is at risk of injury, the buzzer will emit a sharp alarm, the vibration motor will generate a strong vibration, and the red LED will illuminate, alerting the athlete from multiple angles.

[0065] System software architecture: A layered architecture is adopted, including a data acquisition layer, a data transmission layer, a data processing layer, and a user interaction layer. The data acquisition layer is responsible for acquiring raw data from the motion posture detection module and the heart rate monitoring module; the data transmission layer transmits data to the data analysis module via transmission protocols such as Bluetooth; the data processing layer uses machine learning algorithms to analyze and process the data; and the user interaction layer is responsible for presenting warning information to the user.

[0066] Software Implementation of Each Module: The software for the motion posture detection module and heart rate monitoring module controls the hardware devices and acquires data by writing corresponding drivers. Taking the MPU6050 as an example, a driver is written using the I2C communication protocol to set parameters such as the sensor's sampling frequency and measurement range, and to periodically read the data collected by the sensor. The software for the data transmission module utilizes the Bluetooth protocol stack to package, transmit, and receive data. In the nRF52832 Bluetooth module, API functions provided by the Bluetooth protocol stack are used to package the acquired data according to the Bluetooth data packet format, and then transmit it wirelessly to the device where the data analysis module is located. The software for the data analysis module is written in Python and uses machine learning libraries such as Scikit-learn and TensorFlow to implement decision tree algorithms and neural network algorithms. For example, the DecisionTreeClassifier class in the Scikit-learn library is used to build a decision tree model to analyze and judge motion posture and heart rate data; a neural network model is built using the TensorFlow library for training and prediction. The software for the warning module controls the working status of the buzzer, vibration motor, and LED lights based on the output results of the data analysis module. When a sports injury is detected, the buzzer sounds an alarm, the vibration motor starts vibrating, and the LED light illuminates through corresponding control commands.

[0067] Data interaction process: The motion posture detection module and heart rate monitoring module collect data in real time according to the set sampling frequency and send the data to the data transmission module. The data transmission module packages the received data and then transmits it via Bluetooth to the device where the data analysis module is located. After receiving the data, the data analysis module performs data preprocessing, such as data cleaning and normalization, and then uses machine learning algorithms to analyze it and determine if there is any sports injury. If a sports injury is detected, an early warning message is sent to the early warning module, which activates the corresponding early warning device and issues an alert to the user. Simultaneously, the data analysis module can also store the analysis results and historical data in a database for subsequent querying and analysis.

[0068] Hardware and software integration is as follows: The hardware devices for the motion posture detection module, heart rate monitoring module, data transmission module, data analysis module, and early warning module are physically integrated through circuit connections and interface interfaces. For example, the MPU6050 six-axis inertial sensor is connected to the Raspberry Pi 4B's I2C interface, the MAX30102 photoelectric pulse sensor is connected to the Raspberry Pi 4B's SPI interface, and the nRF52832 Bluetooth module is connected to the Raspberry Pi 4B's UART interface. On the software side, the software programs of each module are integrated to ensure smooth data transmission and processing between different modules. The overall system operation is achieved by writing a main program that calls the driver programs and function functions of each module.

[0069] The specific application scenarios are as follows:

[0070] (a) Everyday exercise scenarios

[0071] For ordinary sports enthusiasts, using this system in daily running, fitness, and other activities is simple and easy. Before exercising, first wear the posture detection module and heart rate monitoring module in appropriate positions. The posture detection module can be worn on the wrist, ankle, or waist to ensure accurate collection of posture data from various parts of the body; the heart rate monitoring module is usually worn on the wrist, in close contact with the skin to ensure stable heart rate monitoring. Turn on the system to put it into working mode; the system will automatically initialize settings, including calibrating the sensors and connecting the data transmission module.

[0072] During exercise, the system collects exercise posture and heart rate data in real time. When running, the exercise posture detection module continuously monitors information such as stride frequency, stride length, body tilt angle, and arm swing amplitude; the heart rate monitoring module tracks heart rate changes in real time and transmits this data to the data analysis module via Bluetooth or Wi-Fi. The data analysis module then uses preset algorithms and models to quickly analyze and process the transmitted data to determine if there is a risk of sports injury.

[0073] It is important to ensure the device is securely worn during use to prevent data collection accuracy issues caused by looseness. Before exercising, check that the device has sufficient battery power to ensure the system can operate normally throughout the exercise process. Users should also pay attention to system warnings. If a warning is received, immediately stop exercising, check your physical condition, and take appropriate measures based on the warning prompt. If the warning indicates mild fatigue, reduce the intensity of exercise and take a short rest; if the warning indicates a higher risk of serious sports injury, such as a high risk of joint sprains, seek professional medical help immediately.

[0074] (II) Professional Training Scenarios

[0075] Professional athletes can fully utilize the system's customized features for their specific training needs when using it in training. Before training, coaches or training teams will personalize the system settings based on the athlete's training program, individual physical condition, and training goals. For basketball players, the system will focus on setting parameters related to basketball, such as jump height, landing impact force, joint angle changes during rapid changes of direction, and heart rate threshold ranges under high-intensity exercise.

[0076] During training, the system collects data comprehensively and with high precision. Taking football training as an example, the motion posture detection module not only monitors the athlete's running posture, passing and shooting actions, but also focuses on the body's collision posture and force during confrontations; the heart rate monitoring module tracks the athlete's heart rate changes in real time under different training intensities, including heart rate recovery after rapid sprints, long-distance endurance running, and high-intensity confrontations. The data transmission module quickly and stably transmits this large amount of complex data to the data analysis module.

[0077] The data analysis module utilizes machine learning algorithms and models specifically optimized for professional athlete training to perform in-depth data analysis. By comparing data with historical training data and sports injury cases from a large number of professional athletes, it assesses whether the athlete's current training status poses a potential risk of sports injury. If the system detects that an athlete's joint angle exceeds the normal range during repeated rapid changes of direction, and the heart rate recovery time is excessively long, the system will determine that the athlete may be at risk of joint sprain or muscle over-fatigue, and will promptly issue a warning to the coach and athlete via the alert module.

[0078] Professional athletes should strictly follow the coach's instructions and the system's prompts when using the system. They should promptly report any physical sensations or abnormalities during training to their coach so that the coach can adjust the training plan accordingly based on the system's monitoring data. Coaches should also use the data and warnings provided by the system to scientifically arrange training intensity and rest time to prevent athletes from suffering sports injuries due to overtraining.

[0079] (III) Rehabilitation Training Scenarios

[0080] This system plays a crucial supporting role in the rehabilitation training of injured patients. Before rehabilitation training begins, doctors or rehabilitation therapists will set corresponding monitoring parameters and rehabilitation assessment indicators for the system based on the patient's injury type, severity, and rehabilitation stage. For patients with ankle sprains, the focus will be on the ankle joint's range of motion, muscle strength recovery, and postural changes during walking or simple movements. Simultaneously, a reasonable heart rate range will be set based on the patient's physical condition and rehabilitation progress as a basis for judging whether the intensity of rehabilitation training is appropriate.

[0081] During rehabilitation training, patients wear a motion posture detection module and a heart rate monitoring module to perform corresponding rehabilitation exercises. When performing simple leg flexion and extension exercises, the motion posture detection module accurately monitors the flexion and extension angles of the ankle joint, as well as muscle contraction and relaxation; the heart rate monitoring module records the patient's heart rate changes in real time during training. The data transmission module transmits this data to the data analysis module, which then combines the patient's rehabilitation plan with historical monitoring data to evaluate the effectiveness of the rehabilitation training.

[0082] The system compares current movement posture data with data from the initial rehabilitation phase to determine whether joint range of motion has increased and muscle strength has improved. By analyzing heart rate data, it assesses the patient's physical endurance and recovery progress. If the system finds that joint range of motion improvement is not significant during rehabilitation training, and the heart rate rises abnormally even at low intensity training, it will alert the rehabilitation therapist that there may be slow rehabilitation progress or inappropriate training intensity. The therapist can then adjust the rehabilitation training plan based on the system's suggestions, such as increasing or decreasing training intensity or changing training movements, to promote the patient's recovery. Simultaneously, the system can periodically generate rehabilitation effectiveness evaluation reports, providing objective data support for doctors and rehabilitation therapists to better develop subsequent rehabilitation plans.

[0083] This application also includes a sports injury monitoring method based on digital multi-model assessment of sports injuries, applicable to the aforementioned sports injury monitoring system based on digital multi-model assessment of sports injuries, comprising the following steps: detecting a person's movement posture through a movement posture detection module; monitoring a person's heart rate through a heart rate monitoring module; transmitting the data collected by the movement posture detection module and the heart rate monitoring module through a data transmission module; receiving the transmitted data through a data analysis module and analyzing the data to determine whether a sports injury exists; when the data analysis module determines that a sports injury exists, the early warning module issues an early warning.

[0084] Example 2

[0085] To improve the accuracy and processing speed of sports injury monitoring systems and reduce the failure rate, the motion posture detection module and the heart rate monitoring module can be linked. Data collected from these two modules can yield a dynamic parameter—the Motion Risk Coefficient (MRC).

[0086] The formula for calculating the sports risk factor is as follows:

[0087]

[0088] in:

[0089] - It is real-time acceleration data. It is the average acceleration. It is the standard deviation of acceleration.

[0090] - It is real-time heart rate. It is the average heart rate. It is the standard deviation of heart rate.

[0091] - and It is a weighting coefficient that is adjusted according to the characteristics of different sports and groups of people in order to balance the impact of acceleration and heart rate on sports risks.

[0092] The data analysis module can control the early warning module based on this sports risk coefficient. When the MRC exceeds a preset threshold, the early warning module will issue an alert of the corresponding level, reminding the athlete to be aware of potential sports injury risks. Through such dynamic parameter control, the system can more accurately judge sports injury risks, improve processing speed, and avoid equipment damage or misjudgment due to excessive or untimely warnings, thereby reducing the damage rate. For example, in a high-intensity basketball game, when a player's sports risk coefficient exceeds a set danger threshold, the early warning module will immediately activate multiple warning methods such as sound, vibration, and information push to remind the player and coach to take appropriate measures to ensure the player's safety.

[0093] Example 3

[0094] Current digital sports injury monitoring systems often process data from each module in a fixed or random order during data processing, resulting in insufficient data fusion, delayed identification of abnormal parameters, and accuracy significantly affected by scenario adaptability. Specifically: First, the conventional data processing order (movement posture detection module, data transmission module, data analysis module, heart rate monitoring module, and early warning module) fails to adjust processing priorities based on dynamic data changes during the cycle, leading to untimely extraction of key injury risk parameters (such as abnormality of movement posture and heart rate fluctuation coefficient). Second, fixed early warning thresholds are used, failing to reflect the cumulative characteristics of data through the number of cycles, making it difficult to adapt to the changing injury risk patterns under different exercise intensities (such as high-intensity interval training and long-duration endurance exercise). Third, single data processing is susceptible to instantaneous sensor errors, which cannot be reduced through iterative processing, resulting in insufficient accuracy in injury assessment, especially in special scenarios (such as mountain running and indoor equipment training), where the false positive and false negative rates are high.

[0095] The core of this solution is to construct a data processing mechanism that sequentially cycles through a motion posture detection module, a data transmission module, a data analysis module, a heart rate monitoring module, and an early warning module. The data fusion effect is optimized through iterative iteration, and the early warning threshold is dynamically adjusted based on the number of iterations. The specific technical solution is as follows:

[0096] Motion posture detection module: Employs an inertial measurement unit (IMU) and pressure sensors to collect the athlete's joint angles (such as knee flexion angle and ankle plantar flexion angle) and acceleration (…). Parameters such as triaxial acceleration and ground reaction force are sampled at a frequency according to the formula. ( Sampling frequency, unit The data collection time interval is in seconds (s), representing a typical scenario. Special scenarios can be adaptively adjusted to )set up.

[0097] Heart rate monitoring module: Collects real-time heart rate data via a photoelectric heart rate sensor. (Unit: beats / minute), and calculate the average heart rate during exercise. and heart rate standard deviation Preliminary diagnosis of abnormal heart rate is based on the formula. ( For constants, normal motion High-intensity exercise )implement.

[0098] Early warning module: Receives the analysis results from the data analysis module. When key parameters reach the dynamic threshold, it outputs early warning information through sound and light alarms, mobile APP push notifications, etc. The early warning levels are divided into low risk (prompt to adjust exercise posture), medium risk (suggest reducing exercise intensity), and high risk (forced to stop exercise).

[0099] Data transmission module: Bluetooth The dual-mode transmission protocol prioritizes the transmission of real-time attitude data from the motion attitude detection module via Bluetooth during the loop (transmission rate). (Transmission rate) transmits historical heart rate data from the heart rate monitoring module via Wi-Fi. This ensures that there is no packet loss in different scenarios.

[0100] Data Analysis Module: Extracts posture features from the motion posture detection module using a convolutional neural network (CNN), and analyzes the heart rate temporal features from the heart rate monitoring module using a recurrent neural network (RNN), outputting a key injury risk parameter—the sports injury risk coefficient. The results are then transmitted to the early warning module.

[0101] The cyclic data processing flow is as follows:

[0102] Initialization Phase: After system startup, the motion posture detection module and heart rate monitoring module complete sensor calibration respectively, the data transmission module establishes a communication connection with the data analysis module, the data analysis module loads the pre-trained CNN-RNN fusion model, and the early warning module initializes the early warning threshold. .

[0103] The cyclic processing phase consists of the following steps:

[0104] (1) The motion posture detection module collects motion posture data according to the sampling frequency. Generate pose dataset ;

[0105] (2) The data transmission module will The data is packaged and transmitted to the data analysis module, while simultaneously receiving feedback from the previous analysis results from the data analysis module and adjusting the transmission priority accordingly.

[0106] (3) Data analysis module Feature extraction is performed, and the heart rate monitoring module is triggered to collect real-time heart rate data. Generate heart rate dataset ;

[0107] (4) Integration of data analysis modules and Calculate the risk coefficient of sports injury ;

[0108] (5) Comparison of early warning modules With the current dynamic threshold ,like Then proceed to the next cycle. ;like If the loop stops, the warning data will be output.

[0109] Formula 1: Sports Injury Risk Coefficient Calculation formula ;in, This is the posture weighting coefficient (value range 0.6-0.8, adjusted according to the type of exercise, such as running). ,swim Heart rate weighting coefficient ; The pose is the motion pose anomaly score (output by the CNN model, with a value ranging from 0 to 1). attitude (Indicates a completely abnormal posture); Heart rate is a risk indicator of heart rate fluctuations (output by an RNN model, with a value ranging from 0 to 1). Heart rate This indicates that the heart rate fluctuation is outside the safe range. The specific calculations can be further refined as follows: ,in This refers to the number of attitude parameters (such as knee joint angle, acceleration, etc.). For the first Weights of each parameter These are real-time attitude parameter values. These are the standard attitude parameter values.

[0110] Formula 2: Dynamic Early Warning Threshold Calculation formula ( The number of loops. ): ;in, The initial warning threshold is set (range 0.6-0.8, determined based on the sports population, such as professional athletes). ordinary athletes ); This is the threshold attenuation coefficient (range 0.02–0.05, balancing accuracy and sensitivity, such as in high-intensity exercise). Low-intensity exercise ); Using the natural logarithm function ensures that the threshold decreases slowly with increasing iteration count, improving the ability to identify potential damage risks. When Reaching the maximum number of loops (default When (customizable), if Not yet reached The system will automatically reset the loop. ) and update .

[0111] For example:

[0112] Initialization Phase: After system startup, the motion posture detection module and heart rate monitoring module complete sensor calibration respectively, the data transmission module establishes a communication connection with the data analysis module, the data analysis module loads the pre-trained CNN-RNN fusion model, and the early warning module initializes the early warning threshold. .

[0113] The cyclic processing phase consists of the following steps:

[0114] (1) The motion posture detection module collects motion posture data according to the sampling frequency. Generate pose dataset It includes five parameters such as knee joint angle and ankle joint acceleration. ;

[0115] (2) The data transmission module will The data is packaged and transmitted to the data analysis module, while simultaneously receiving feedback from the previous analysis results from the data analysis module and adjusting the transmission priority accordingly.

[0116] (3) Data analysis module Feature extraction is performed, and the heart rate monitoring module is triggered to collect real-time heart rate data. Generate heart rate dataset Suppose a heart rate data collection session was... times / minute;

[0117] (4) Integration of data analysis modules and Calculate the risk coefficient of sports injury ;

[0118] (5) Comparison of early warning modules With the current dynamic threshold ,like Then proceed to the next cycle. ;like If the loop stops, the warning data will be output.

[0119] Formula 1: Sports Injury Risk Coefficient Calculation formula

[0120] ;

[0121] Taking running as an example, ; Assumption The pose is output by the CNN model. If the RNN model outputs 0.3, then: ;

[0122] The specific calculations can be further refined as follows: Assuming Weights of each parameter Real-time knee joint angle Standard value Real-time ankle acceleration value Standard value (If only 2 parameters are listed), then: ;

[0123] Formula 2: Dynamic Early Warning Threshold Calculation formula ( The number of loops. )

[0124] Assuming it is aimed at ordinary athletes, For low-intensity exercise, .when During the next iteration:

[0125]

[0126] when Reaching the maximum number of loops ,like Not yet reached The system will automatically reset the loop. ) and update .

[0127] The working principle is as follows:

[0128] After the system starts, the motion posture detection module first collects initial posture data through the IMU sensor, triggering the data transmission module to transmit the data to the data analysis module using a dual-mode protocol. Upon receiving the posture data, the data analysis module starts a CNN model for feature extraction and simultaneously sends a trigger signal to the heart rate monitoring module to collect real-time heart rate data and transmit it to the data analysis module. The data analysis module analyzes the temporal characteristics of heart rate using an RNN model and, combined with the posture features, calculates the sports injury risk coefficient according to Formula 1. The data analysis module will then... The data is transmitted to the early warning module (early warning), which calculates the current cycle count according to Formula 2. Corresponding dynamic threshold ,contrast and :like The module sends a loop continuation signal, triggering the next loop process; if The warning module immediately stops looping and outputs... Values, abnormal parameters (such as abnormal posture parameters, heart rate values), number of cycles Collect the data and activate the corresponding level of early warning.

[0129] In mountain running scenarios, the motion posture detection module automatically sets the sampling time interval. Adjusted to 0.005s (sampling frequency) This enhances the ability to capture sudden attitude changes caused by rough terrain; the data transmission module prioritizes switching to Wi-Fi transmission (with stronger anti-interference capabilities) to ensure stable data transmission; the data analysis module dynamically adjusts... It is 0.8 (with a higher attitude weight). The threshold decays faster (to 0.04); in indoor yoga scenarios, the motion posture detection module... Adjusted to It is 0.6 (heart rate has a higher weight). The threshold decays more slowly, and the parameter adaptation is automatically completed through a scene recognition algorithm (based on posture and heart rate data within the initial 10 seconds).

[0130] In each loop, the data analysis module will take the data from the previous loop... The value is compared with the actual movement status (such as whether there is any feedback of slight discomfort) and then corrected. and The weighting coefficients; simultaneously, the data transmission module uses data verification algorithms (such as CRC check) to eliminate abnormal data during transmission, ensuring... and The integrity of the sensor is ensured; the motion posture detection module and heart rate monitoring module perform sensor drift correction during cycle intervals (such as zero-drift calibration based on the static state) to reduce the impact of accumulated errors. The impact of value calculation.

[0131] Through a cyclic data processing mechanism, the system iteratively fuses motion posture and heart rate data, and effectively reduces the impact of errors from a single data dimension by adjusting dynamic weighting coefficients. The dynamic warning threshold optimizes with each iteration, avoiding missed detections caused by fixed thresholds and enabling more accurate capture of potential injury risks. The system features adaptive parameter adjustment for different sports scenarios (mountain, indoor, water, etc.), automatically optimizing key parameters such as sampling frequency, data transmission protocol, and weighting coefficients based on scenario characteristics. This overcomes the "one-size-fits-all" processing shortcomings of traditional systems, ensuring stable operation in various complex scenarios. The cyclic iterative processing mechanism effectively offsets instantaneous fluctuations in sensors through multiple data acquisitions and fusions. Combined with sensor drift correction and data verification algorithms, it reduces the impact of accumulated errors on injury risk calculation, providing a reliable data foundation for accurate judgment. The dynamic warning threshold decreases slowly with each iteration, allowing the system to trigger warnings promptly as injury risks gradually accumulate. Compared to traditional fixed threshold alarms, this allows for earlier identification of potential risks, providing athletes with sufficient adjustment time and effectively reducing the incidence of acute injuries.

[0132] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0133] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A sports injury monitoring system for assessing sports injuries based on digitized multi-model state, characterized in that, include: The motion posture detection module is used to detect a person's motion posture; Heart rate monitoring module, used to monitor a person's heart rate; The data transmission module is used to transmit the data collected by the motion posture detection module and the heart rate monitoring module; The data analysis module is used to receive the transmitted data and analyze it to determine whether there is any sports injury. The early warning module issues an early warning when the data analysis module determines that a sports injury exists.

2. The sports injury monitoring system based on digital multimodal assessment of sports injuries according to claim 1, characterized in that, The motion posture detection module adopts one or more combinations of an acceleration sensor, a gyroscope and a magnetometer; the data collection frequency of the sensor satisfies the formula: wherein, is the collection time interval; the motion posture parameters that the motion posture detection module can detect include a motion trajectory, a joint angle and a body posture angle.

3. The sports injury monitoring system based on digital multimodal assessment of sports injuries according to claim 1, characterized in that, The heart rate monitoring module uses a photoelectric pulse sensor; the judgment of abnormal heart rate is based on a formula. ,in, For real-time heart rate, This is the average heart rate. It is a constant. The standard deviation is given; the heart rate monitoring module filters the collected heart rate data to remove noise interference.

4. The sports injury monitoring system based on digital multimodal assessment of sports injuries according to claim 1, characterized in that, The data transmission module employs one or more transmission protocols, including Bluetooth, Wi-Fi, and ZigBee; and in different scenarios, it ensures transmission rate by adaptively adjusting transmission power, channel selection, and data packet segmentation strategies.

5. The sports injury monitoring system based on digital multimodal assessment of sports injuries according to claim 1, characterized in that, The data analysis module uses one or more of the following machine learning algorithms: decision tree algorithm, support vector machine algorithm, and neural network algorithm, to fuse and analyze the multi-source data collected by the motion posture detection module and the heart rate monitoring module to determine the type and degree of sports injury.

6. The sports injury monitoring system based on digital multimodal assessment of sports injuries according to claim 1, characterized in that, The warning module provides warnings in one or more of the following ways: sound warning, vibration warning, light warning, and information push warning. The warning threshold is set based on historical sports injury data of different sports and sports groups, as well as professional sports medicine advice, and the warning threshold can be adjusted according to actual usage.

7. The sports injury monitoring system based on digital multimodal assessment of sports injuries according to any one of claims 6, characterized in that, The motion posture detection module, heart rate monitoring module, data transmission module, data analysis module, and early warning module work together. The motion posture detection module and heart rate monitoring module transmit the collected data to the data analysis module through the data transmission module. After the data analysis module analyzes and processes the data, if it determines that there is a sports injury, it controls the early warning module to issue an early warning, so as to realize the system's rapid and accurate monitoring of sports injuries and improve the safety and sports experience of athletes.

8. A sports injury monitoring system based on digital multi-model assessment of sports injuries as described in claim 7, characterized in that: The motion posture detection module and the heart rate monitoring module are linked to obtain a dynamic parameter—the exercise risk coefficient. The formula for calculating the exercise risk coefficient is as follows: in: It is real-time acceleration data. It is the average acceleration. It is the standard deviation of acceleration. It is real-time heart rate. It is the average heart rate. It is the standard deviation of heart rate. and It is a weighting coefficient that is adjusted according to the characteristics of different sports and groups of people in order to balance the impact of acceleration and heart rate on sports risks.

9. A sports injury monitoring system based on digital multi-model assessment of sports injuries as described in claim 8, characterized in that: The system includes a loop module, and the processing flow of the loop module includes: The motion posture detection module collects motion posture data and analyzes it according to the sampling frequency. Generate pose dataset ; The data transmission module will The data is packaged and transmitted to the data analysis module, while simultaneously receiving feedback from the previous analysis results from the data analysis module and adjusting the transmission priority accordingly. Data analysis module Feature extraction is performed, and the heart rate monitoring module is triggered to collect real-time heart rate data. Generate heart rate dataset ; Data analysis module integration and Calculate the risk coefficient of sports injury ; Comparison of early warning modules With the current dynamic threshold ,like Then proceed to the next cycle. ;like If the loop stops, the warning data will be output.

10. A sports injury monitoring method based on digital multi-model assessment of sports injuries, characterized in that, The sports injury monitoring system based on digital multi-model assessment of sports injuries according to any one of claims 1-9 includes the following steps: The motion posture of a person is detected through a motion posture detection module; The heart rate is monitored using a heart rate monitoring module. The data collected by the motion posture detection module and the heart rate monitoring module are transmitted through the data transmission module; The data analysis module receives and analyzes the transmitted data to determine whether there is any sports injury. When the data analysis module determines that there is a sports injury, the early warning module issues an early warning.

Images