Baduanjin training resistance elastic band and muscle strength evaluation system

Abstract

The application relates to an anti-resistance elastic band for Ba Duan Jin training and a muscle strength evaluation system, which are applied to the technical field of sports training monitoring and evaluation. The system comprises the following: the anti-resistance elastic band is internally integrated with a multi-point flexible film type tension sensor array, which is embedded according to preset layout rules according to the distribution of main force muscle groups of eight Ba Duan Jin postures; a signal conditioning circuit amplifies and filters sensor signals; a data acquisition module receives real-time tension data; a storage stores standard mechanical models of each posture of Ba Duan Jin; a processor executes a program to perform posture recognition, data segmentation, and calculate strength time matching degree, strength amplitude matching degree and strength duration matching degree, and generate multi-dimensional muscle strength evaluation results; and a feedback module outputs the evaluation results. Through special sensor layout, offset detection compensation, elastic coefficient self-calibration and multi-modal fusion, high-precision, long-term stable quantitative evaluation of muscle strength in Ba Duan Jin training is realized.

Description

Technical Field

[0001] This invention relates to the field of sports training monitoring and evaluation technology, specifically to a resistance elastic band and muscle strength assessment system for Baduanjin training. Background Technology

[0002] Baduanjin, a traditional Chinese exercise for health and fitness, consists of eight movements and is believed to regulate the internal organs, unblock meridians, and strengthen the body. With the advancement of national fitness and the Healthy China strategy, the popularity of Baduanjin is constantly increasing, and practitioners' demand for scientific training guidance is also growing. How to quantitatively assess muscle strength during Baduanjin training has become a research hotspot in the field of sports science.

[0003] In existing technologies, motion assessment mainly adopts the following two technical approaches: One type is motion capture and analysis technology based on images / videos: it captures images of the user's movement through a camera, uses computer vision technology to identify the posture of the movement, and evaluates the standardization of the movement. This type of method can provide intuitive visual feedback, but it is greatly affected by factors such as ambient lighting, shooting angle, and occlusion, and it cannot directly measure muscle force data, making it difficult to achieve quantitative assessment of muscle strength.

[0004] Another type is force measurement technology based on wearable sensors: by integrating force sensors onto a carrier such as an elastic band, mechanical data is collected during movement. This type of method can directly measure force parameters, providing new possibilities for muscle strength assessment.

[0005] However, when the above technology is applied to the specific fitness method of Baduanjin, the following shortcomings still exist: the layout of existing elastic band sensors is mostly a general design and has not been optimized for the force characteristics of the eight movements of Baduanjin, resulting in insufficient targeting and accuracy of data collection; at the same time, during use, factors such as wear displacement, material aging, and temperature changes can easily cause sensor position shift and elastic coefficient drift, affecting the consistency of data over long-term use.

[0006] Therefore, how to provide a solution specifically designed for Baduanjin (Eight Pieces of Brocade) that can achieve accurate and stable muscle strength assessment is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention aims to provide a resistance band and muscle strength assessment system for Baduanjin (Eight Pieces of Brocade) training. First, by employing a specialized sensor layout tailored to the eight movements of Baduanjin, precise force data acquisition is achieved, overcoming the limitations of existing universal and specific elastic band sensor layouts. Furthermore, sensor position offset detection compensation and elastic coefficient self-calibration ensure long-term data stability. Simultaneously, multimodal fusion recognition and multi-dimensional evaluation indicators are introduced to improve the accuracy of movement recognition and the comprehensiveness of the assessment. Ultimately, this achieves high-precision, long-term stable quantitative assessment of muscle strength during Baduanjin training.

[0008] To achieve the above objectives, in a first aspect, the present invention provides a muscle strength assessment system for Baduanjin training, comprising: A resistance elastic band, comprising an elastic band body and a multi-point flexible thin-film tension sensor array integrated inside the elastic band body; The multi-point flexible thin-film tension sensor array is embedded at the key stress points of the elastic band body according to the distribution of the main force-generating muscle groups corresponding to the eight moves of Baduanjin, and according to the preset sensor layout rules. The sensor layout rules include the spacing between the sensors, the layout density, and the way they are combined with the elastic band body. The resistance elastic band also includes a signal conditioning circuit, which is electrically connected to a multi-point flexible thin-film tension sensor array and integrated inside the elastic band body or at the fixed end of the elastic band body. It is used to amplify and filter the raw signals collected by the sensors. The data acquisition module is connected in communication with the signal conditioning circuit and is used to receive and process the amplified and filtered real-time tension data. The memory stores the standard mechanical models of each move of Baduanjin. The standard mechanical models are constructed based on the time-tension characteristic data collected from the standard movement samples of Baduanjin. They include the preset tension thresholds at the start time, peak time, and end time for each move of Baduanjin, as well as the rising slope characteristic curve from the start to the peak time and the falling slope characteristic curve from the peak time to the end time. When a processor executes a computer program stored in memory, it performs the following steps: Receive real-time tension data from the data acquisition module; The standard mechanical models of each move of Baduanjin stored in the memory are called to identify the move in real-time tension data in order to determine the specific move of Baduanjin that the user is currently performing. Based on the temporal characteristics of the standard mechanical model corresponding to the identified specific move, the real-time tension data is automatically divided into independent data segments corresponding to that move; For each independent data segment, its matching degree with the corresponding standard mechanical model is calculated to generate multi-dimensional muscle strength assessment results, which include strength temporal matching degree, strength amplitude matching degree, and strength duration matching degree. The feedback module is used to output the multi-dimensional muscle strength assessment results in a way that is perceptible to the user.

[0009] As a further improvement of the present invention, the sensors of the multi-point flexible thin film tension sensor array adopt a distributed layout, and the spacing between adjacent sensors is a preset spacing determined according to the concentration of the force points of the corresponding moves. The layout density is dynamically adjusted according to the concentration of the force points of the corresponding moves.

[0010] As a further improvement to the present invention, the sensor layout rule is specifically as follows: For the "Holding up the sky with both hands to regulate the three jiaos" move in Baduanjin, the first sensor group is set at the key stress points of the corresponding shoulder and back muscle groups. To counter the "drawing the bow with both hands like shooting an eagle" move, a second sensor group was installed at the key stress points of the corresponding chest, back, upper limb and lower limb muscle groups. For the "single lift to regulate the spleen and stomach" move, a third sensor group is set at the key stress points of the corresponding waist, abdomen and upper limb muscle groups. To address the "five strains and seven injuries from looking backward" technique, a fourth sensor group was installed at the key stress points corresponding to the neck, shoulder, back, and upper limb muscle groups. To counter the "shaking head and wagging tail to extinguish heart fire" move, a fifth sensor group is installed at the key stress points corresponding to the neck, waist, hips and lower limb muscle groups. For the "Two Hands Climbing Feet to Strengthen Kidneys and Waist" move, a sixth sensor group is set at the key stress points of the corresponding back and lower limb muscles. To address the "clenched fist and glaring eyes to increase strength" move, a seventh sensor group was installed at the key stress points of the corresponding upper limb, shoulder and back, and lower limb muscle groups. To address the "Seven Bounces on the Back to Cure All Diseases" technique, an eighth sensor group was installed at key stress points on the corresponding lower limbs and foot muscles.

[0011] As a further improvement of the present invention, the processor also performs the following steps: Based on real-time tension data collected by a multi-point tension sensor array, the signal correlation between adjacent sensors is calculated. When the correlation between adjacent sensor signals is detected to be lower than a preset threshold, it is determined that a sensor position shift has occurred. In response to the judgment, a calibration prompt signal is issued, or an offset compensation algorithm is automatically executed, which includes re-establishing the mapping relationship between the sensor and the key force points.

[0012] As a further improvement of the present invention, the system also includes an inertial measurement unit, integrated on a resistance elastic band, for synchronously collecting the user's motion posture data during Baduanjin training. The processor also performs time synchronization and feature layer fusion of real-time tension data and motion posture data, inputting them into a pre-trained multimodal motion recognition model to identify the specific move of Baduanjin that the user is currently performing; Among them, the multimodal motion recognition model is a neural network model trained based on sample data, and its inputs include tension feature vectors and posture feature vectors.

[0013] As a further improvement of the present invention, the system also includes an elastic coefficient self-calibration unit, which is configured as follows: Before training begins or during training intervals, the system receives calibration tension data collected by a multi-point flexible thin-film tension sensor array when the user performs preset standard calibration actions. The calibration tension data is compared with the reference tension data corresponding to the standard calibration action stored in the memory, and the current elastic coefficient correction factor is calculated. The correction factor is applied to the correction of subsequent real-time tension data to eliminate tension data drift caused by elastic band aging or temperature changes.

[0014] As a further improvement to the present invention, the multidimensional muscle strength assessment results include: Force timing matching degree is used to characterize the degree of consistency between the user's actual force application timing and the preset force application timing in the standard mechanical model; Force amplitude matching degree is used to characterize the percentage of the user's actual force amplitude compared to the preset force amplitude in the standard mechanical model; Strength consistency matching is used to characterize the degree to which the tension decay rate of the user during the hold phase matches the preset decay rate in the standard mechanical model.

[0015] Secondly, the present invention provides a resistance elastic band for Baduanjin training, comprising: Elastic band body; The multi-point flexible thin-film tension sensor array consists of multiple flexible thin-film tension sensors arranged according to the distribution of the main force-generating muscle groups corresponding to the eight moves of Baduanjin. Each sensor is embedded inside the elastic band body according to the preset sensor layout rules, which include the spacing between sensors, the layout density, and the way they are connected to the elastic band body. The signal conditioning circuit is electrically connected to the multi-point flexible thin-film tension sensor array and is integrated inside the elastic band body or at the fixed end of the elastic band body. It is used to amplify and filter the raw signals collected by the sensors. The data interface, electrically connected to the signal conditioning circuit, is used to output the amplified and filtered real-time tension data.

[0016] Thirdly, the present invention provides a muscle strength assessment method based on the above-mentioned system, comprising the following steps: By integrating a multi-point flexible thin-film tension sensor array inside the resistance elastic band, the original tension signals of multiple key force points are simultaneously collected when the user is performing Baduanjin training. By using a signal conditioning circuit integrated on the resistance elastic band, the original tension signal is amplified and filtered to obtain real-time tension data. The system calls up the pre-stored standard mechanical models of each move in Baduanjin and identifies the moves in real-time tension data to determine the specific move of Baduanjin that the user is currently performing. The standard mechanical model includes the tension thresholds at the start time, peak time, and end time for each move in Baduanjin, as well as the rising slope characteristic curve from the start to the peak and the falling slope characteristic curve from the peak to the end. Based on the temporal characteristics of the standard mechanical model corresponding to the identified specific move, the real-time tension data is automatically divided into independent data segments corresponding to that move; For each independent data segment, its matching degree with the corresponding standard mechanical model is calculated to generate multi-dimensional muscle strength assessment results. The multi-dimensional muscle strength assessment results include force timing matching degree, force amplitude matching degree, force duration matching degree, and bilateral coordination coefficient. Among them, the bilateral coordination coefficient is based on the time series data of the corresponding sensors on the left and right sides collected by the multi-point tension sensor array, and is obtained by calculating the synchronicity and amplitude ratio of the force exertion sequence on the left and right sides. The multi-dimensional muscle strength assessment results are output to the user through the feedback module.

[0017] As a further improvement of the present invention, the above method also includes an elastic coefficient self-calibration step: Before training begins or during training intervals, the system receives calibration tension data collected by a multi-point flexible thin-film tension sensor array when the user performs preset standard calibration actions. The calibration tension data is compared with the reference tension data corresponding to the pre-stored standard calibration action to calculate the current elastic coefficient correction factor; The correction factor is applied to the correction of subsequent real-time tension data to eliminate tension data drift caused by elastic band aging or temperature changes.

[0018] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention achieves precise capture of the force characteristics of the eight movements of Baduanjin (Eight Pieces of Brocade) through a dedicated sensor layout rule. The sensor group layout for each movement is determined based on the biomechanical characteristics of that movement, avoiding data redundancy and loss of key information caused by general sensor layouts, and significantly improving the targeting and accuracy of the assessment.

[0019] 2. This invention, through correlation analysis of adjacent sensor signals, can detect in real time the sensor position shift caused by the elastic band's displacement during use, and correct it through calibration prompts or automatic compensation algorithms. By detecting displacement in real time through correlation analysis of adjacent sensor signals and correcting it through calibration prompts or automatic compensation algorithms, this invention solves the problem of data inaccuracy caused by wearing displacement in existing elastic band sensors, ensuring data reliability for long-term use and reducing the inconvenience of users frequently adjusting the elastic band.

[0020] 3. This invention utilizes an elastic coefficient self-calibration unit to periodically calibrate the elastic coefficient using standard calibration actions, effectively eliminating tension data drift caused by elastic band material aging and environmental temperature changes. This extends the product's effective service life and effectively ensures the comparability and consistency of long-term training data.

[0021] 4. This invention overcomes the low accuracy of single-modal data recognition by fusing inertial measurement units with tension data, combining mechanical features (tension) with motion features (posture). The use of a neural network model for feature layer fusion significantly improves performance, particularly in recognizing complex moves, laying the foundation for subsequent accurate evaluation.

[0022] 5. This invention, through a bilateral coordination coefficient based on time-series data from multi-point tension sensors, calculates the temporal synchronicity and dynamic amplitude ratio of force exerted on both sides, thus achieving a quantitative assessment of the degree of bilateral muscle group coordination in symmetrical / asymmetrical movements of Baduanjin (Eight Pieces of Brocade). Unlike existing symmetry assessment methods based on static force comparison, the bilateral coordination coefficient of this invention, based on time-series data, reflects the dynamic coordination during the movement, providing a more scientific basis for targeted training.

[0023] 6. This invention comprehensively evaluates a user's muscle strength performance from multiple dimensions, such as the timing of force exertion, the magnitude of force exertion, and the ability to maintain strength, by matching the strength timing, the magnitude of force exertion, and the strength duration. The evaluation dimensions are comprehensive and the information is rich, which can provide users with more valuable training guidance. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the resistance elastic band used for Baduanjin training in the first embodiment of the present invention; Figure 2This is a structural block diagram of the muscle strength assessment system for Baduanjin training in the second embodiment of the present invention; Figure 3 This is a flowchart of the muscle strength assessment method in the third embodiment of the present invention. Detailed Implementation

[0025] The three embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0026] First implementation method: Figure 1 This invention illustrates a resistance elastic band for Baduanjin (Eight Pieces of Brocade) training, which includes an elastic band body, a multi-point flexible thin-film tension sensor array, a signal conditioning circuit, and a data interface.

[0027] The elastic band body is made of thermoplastic polyurethane elastomer material, with the following characteristics: Shore hardness 85A, tensile strength 35MPa, elongation at break 500%, and working temperature -20℃~60℃. The elastic band body length is adjustable, ranging from 80 to 140cm, with a width of 10cm and a thickness of 2.5mm. The elastic band has snap-on fastening ends at both ends, one end being a pin structure and the other a socket structure, allowing for quick on and off.

[0028] The multi-point flexible thin-film tension sensor array consists of multiple flexible thin-film tension sensors arranged according to the distribution of the main force-generating muscle groups corresponding to the eight movements of Baduanjin (Eight Pieces of Brocade). Each sensor is embedded inside the elastic band body according to a preset sensor layout rule. The sensor layout rule includes the spacing between sensors, the layout density, and the way they are connected to the elastic band body.

[0029] The spacing between adjacent sensors is a preset spacing determined based on the concentration of force points for the corresponding moves. For example, for moves with concentrated force points (such as the upper arm muscle area in "clenching fists and glaring to increase strength"), the preset spacing is 2-3 cm; for moves with dispersed force points (such as the shoulder and back area in "holding up the sky with both hands to regulate the three jiaos"), the preset spacing is 4-5 cm; and for moves with moderately concentrated force points, the preset spacing is 3-4 cm. These numerical ranges were determined by statistically analyzing the diameter of the force point distribution areas of 50 test subjects, ensuring that a single sensor can cover the main force-generating areas. The layout density is dynamically adjusted according to the concentration of force points for the corresponding moves: based on the identified current move, the processor automatically calls upon the corresponding subset of sensors for focused sampling, reducing the sampling frequency of sensors in non-critical areas, thus optimizing the allocation of computing resources.

[0030] The specific sensor layout rules are as follows: For the "Holding up the sky with both hands to regulate the three jiaos" move in Baduanjin, the first sensor group is set at the key stress points of the corresponding shoulder and back muscle groups. To counter the "drawing the bow with both hands like shooting an eagle" move, a second sensor group was installed at the key stress points of the corresponding chest, back, upper limb and lower limb muscle groups. For the "single lift to regulate the spleen and stomach" move, a third sensor group is set at the key stress points of the corresponding waist, abdomen and upper limb muscle groups. To address the "five strains and seven injuries from looking backward" technique, a fourth sensor group was installed at the key stress points corresponding to the neck, shoulder, back, and upper limb muscle groups. To counter the "shaking head and wagging tail to extinguish heart fire" move, a fifth sensor group is installed at the key stress points corresponding to the neck, waist, hips and lower limb muscle groups. For the "Two Hands Climbing Feet to Strengthen Kidneys and Waist" move, a sixth sensor group is set at the key stress points of the corresponding back and lower limb muscles. To address the "clenched fist and glaring eyes to increase strength" move, a seventh sensor group was installed at the key stress points of the corresponding upper limb, shoulder and back, and lower limb muscle groups. To address the "Seven Bounces on the Back to Cure All Diseases" technique, an eighth sensor group was installed at key stress points on the corresponding lower limbs and foot muscles.

[0031] The specific locations of each sensor group (based on human anatomical landmarks) are as follows: First sensor group: Three sensors are set up, one on the left and one on the right, located 5cm inside the acromion and at the midpoint of the scapular spine.

[0032] The second sensor group consists of six sensors located on the lateral side of the pectoralis major (3cm below the midpoint of the clavicle), the upper border of the latissimus dorsi (at the level of the inferior angle of the scapula), the middle of the biceps brachii, the middle of the triceps brachii, the middle of the quadriceps femoris (10cm above the patella), and the middle of the hamstrings.

[0033] The third sensor group consists of three sensors located in the upper part of the rectus abdominis muscle (3cm below the xiphoid process), the lower part of the rectus abdominis muscle (3cm below the umbilicus), and the middle part of the forearm extensor muscles (5cm above the ulnar styloid process).

[0034] The fourth sensor group consists of four sensors located in the splenius capitis muscle (2 cm lateral to the second cervical vertebra), levator scapulae (superior angle of the scapula), middle trapezius muscle (midpoint of the line connecting the seventh cervical vertebra and the acromion), and middle forearm flexor muscles (5 cm above the styloid process of the ulna).

[0035] The fifth sensor group consists of five sensors located in the splenius muscle of the neck (2 cm lateral to the second cervical vertebra), the middle of the gluteus maximus, the middle of the quadriceps femoris (10 cm above the patella), the middle of the hamstrings, and the middle of the gastrocnemius muscle.

[0036] The sixth sensor group consists of three sensors located at the lower segment of the erector spinae muscle (3cm lateral to the fourth lumbar vertebra), the lower segment of the hamstring muscle, and 5cm above the Achilles tendon.

[0037] The seventh sensor group consists of six sensors located in the middle of the biceps brachii, the middle of the triceps brachii, the anterior deltoid, the middle of the forearm flexor muscles, the middle of the quadriceps femoris (10cm above the patella), and the middle of the gastrocnemius muscle.

[0038] The eighth sensor group consists of three sensors located in the middle of the gastrocnemius muscle, the middle of the soleus muscle, and the corresponding position of the Yongquan acupoint on the sole of the foot.

[0039] The above layout rules were determined based on the following: surface electromyography (EMG) signals and pressure distribution data were collected from 50 professional Baduanjin coaches or experienced practitioners with over 5 years of practice experience. The distribution areas of the main force-generating muscle groups under each movement were statistically analyzed to identify the locations of the muscle groups that exert the most significant force and contribute the most to the completion of the movement, serving as key pressure points. Specifically, "main force-generating muscle groups" refer to muscle groups that play a decisive role in completing a specific Baduanjin movement, identified by EMG signal amplitudes exceeding 40% of the maximum voluntary contraction amplitude; "key pressure points" refer to the points in the contact area between the elastic band and the body where tension changes most significantly, identified by the center point of the area where the pressure value exceeds 30% of the peak pressure in the pressure distribution test.

[0040] The signal conditioning circuit is electrically connected to a multi-point flexible thin-film tension sensor array and is integrated inside the elastic band body or at a fixed end of the elastic band body. It is used to amplify and filter the raw signals acquired by the sensors. The signal conditioning circuit includes a multi-channel instrumentation amplifier array, an 8th-order Butterworth active low-pass filter (cutoff frequency 10Hz), a 12-bit analog-to-digital converter, and a microcontroller. The microcontroller is responsible for timing control, data packaging, and communication.

[0041] The data interface is electrically connected to the signal conditioning circuit and is used to output the amplified and filtered real-time tension data. The data interface supports both wired and wireless output methods: the wired interface uses a USB-C interface and supports USB 2.0 high-speed transmission; the wireless interface uses a Bluetooth 5.0 module and supports wireless connection with smartphones, tablets, and other devices.

[0042] The elastic band itself also integrates an inertial measurement unit (IMU) for synchronously collecting motion posture data of the user during Baduanjin training. The IMU includes a three-axis accelerometer and a three-axis gyroscope, with a sampling frequency of 100Hz, consistent with the sampling frequency of the tension sensor array, and time synchronization is achieved through a unified hardware clock trigger signal.

[0043] The elastic band itself integrates an elastic coefficient self-calibration unit. This unit is configured to: receive calibration tension data collected by a multi-point flexible thin-film tension sensor array when the user performs a preset standard calibration action before training begins or during training intervals; compare the calibration tension data with the baseline tension data corresponding to the standard calibration action pre-stored in the memory, calculate the current elastic coefficient correction factor; and apply the correction factor to the correction of subsequent real-time tension data to eliminate tension data drift caused by elastic band aging or temperature changes. The standard calibration action is designed as extending both arms horizontally forward, shoulder-width apart, and holding for 5 seconds. The baseline tension data is obtained before the elastic band leaves the factory: at a standard ambient temperature of 25℃, 10 test subjects with standard body types perform the standard calibration action, collect tension data from each sensor, and take the average value as the baseline tension data stored in the memory. The correction factor calculation formula is: (Calculation per sensor); in This indicates the sensor number. In subsequent use, the real-time tension data Frawi(t) is multiplied by a correction factor. Perform correction: ; Optionally, a digital temperature sensor can be integrated into the elastic band to establish a temperature-correction factor mapping table and achieve real-time temperature compensation.

[0044] Second implementation method: Figure 2 A muscle strength assessment system for Baduanjin training is shown, which includes the resistance elastic band described in the first embodiment, as well as a data acquisition module, a memory, a processor, and a feedback module.

[0045] The data acquisition module communicates with the signal conditioning circuit on the resistance elastic band to receive and process amplified and filtered real-time tension data. The data acquisition module can also use Bluetooth communication to wirelessly connect with the resistance elastic band and send the real-time tension data to a smartphone app.

[0046] The memory stores standard mechanical models for each movement of Baduanjin. These models are constructed based on time-tension characteristic data collected from standard Baduanjin movement samples. They include preset tension thresholds for the start, peak, and end times for each movement, as well as the upward slope characteristic curve from start to peak and the downward slope characteristic curve from peak to end. The construction method for the standard mechanical models is as follows: Standard movement samples from 50 professional Baduanjin instructors were collected, with each movement repeated 10 times, and tension data from multiple channels were recorded simultaneously. After preprocessing the raw data, a dynamic time warping algorithm was used to align the time-series data of each sample. For each Baduanjin movement, the start time (the moment when tension rises to 10% of the peak), peak time, end time (the moment when tension falls to 10% of the peak), upward slope (the average rate of rise from start to peak), and downward slope (the average rate of fall from peak to end) were extracted. The mean and standard deviation of each feature were calculated to construct a multi-dimensional standard mechanical model.

[0047] When a processor executes a computer program stored in memory, it performs the following steps: Receive real-time tension data from the data acquisition module; The standard mechanical models of each move of Baduanjin stored in the memory are called to identify the move in real-time tension data in order to determine the specific move of Baduanjin that the user is currently performing. Based on the temporal characteristics of the standard mechanical model corresponding to the identified specific move, the real-time tension data is automatically divided into independent data segments corresponding to that move; For each independent data segment, its matching degree with the corresponding standard mechanical model is calculated to generate multi-dimensional muscle strength assessment results, which include strength temporal matching degree, strength amplitude matching degree, and strength duration matching degree.

[0048] The move recognition process employs a dynamic time warping algorithm, which matches the real-time tension data sequence with the standard mechanical models of each move based on similarity, selecting the move with the smallest similarity distance as the recognition result. To improve recognition accuracy, a sliding time window can be used for continuous recognition, taking the majority vote result.

[0049] Data segmentation is based on the temporal characteristics of the standard mechanics model corresponding to the identified moves, detecting the start and end boundaries of the moves in real-time data. The start boundary detection method is as follows: the time point when the tension data first exceeds 80% of the start time tension threshold is detected, and the time point when the tension is 50% lower than the start time threshold is taken as the start boundary; the end boundary detection method is as follows: the time point when the tension data last falls below 120% of the end time tension threshold is detected, and the time point when the tension is 50% lower than the end time threshold is taken as the end boundary.

[0050] Matching degree calculation includes: Force timing matching degree: Characterizes the degree of consistency between the user's actual force application timing and the preset force application timing in the standard mechanical model. The calculation formula is: ; in, The moment when the user actually exerts force to reach its peak (relative to the starting point of the move). This is the standard peak time.

[0051] Force amplitude matching degree: Characterizes the percentage of the user's actual force amplitude compared to the preset force amplitude in the standard mechanical model. The calculation formula is: in, The peak tension is the actual force exerted by the user, while the standard peak tension is the standard peak tension.

[0053] Strength Consistency Match: Characterizes the degree to which the user's tension decay rate during the hold phase matches the preset decay rate in the standard mechanics model. The calculation formula is: ; That This represents the average descent slope from the peak to the end point in the user's actual data. The standard descent slope.

[0054] The feedback module is used to output the multi-dimensional muscle strength assessment results in a user-perceptible manner, including a display screen, speaker, and vibrator, to provide feedback to the user in the form of visual reports, voice prompts, or tactile signals.

[0055] The processor also performs sensor position offset detection and compensation steps: Based on real-time tension data collected by a multi-point tension sensor array, the signal correlation between adjacent sensors is calculated. When the correlation between adjacent sensor signals is detected to be lower than a preset threshold, it is determined that a sensor position shift has occurred. In response to the judgment, a calibration prompt signal is issued, or an offset compensation algorithm is automatically executed, which includes re-establishing the mapping relationship between the sensor and the key force points.

[0056] Signal correlation was calculated using the Pearson correlation coefficient, with a sliding time window of 1 second (100 sampling points), sliding once every 0.5 seconds. The formula for the Pearson correlation coefficient is: ; Under normal circumstances, the correlation between adjacent sensor signals should be high (>0.8); when the correlation is lower than 0.6 and lasts for more than 3 seconds, a sensor position shift is determined to have occurred. The automatic compensation mode of the shift compensation algorithm includes: based on the signal distribution of the current sensor array, the sensors are re-divided into 8 functional groups using the K-means clustering algorithm; the re-divided functional groups are matched with the eight moves of Baduanjin (based on the correlation between the signals of each functional group and the standard mechanical model of each move); the sensor weights in the move recognition model are updated, and a new mapping relationship between sensors and key force points is established.

[0057] The processor also performs a multimodal motion recognition step: real-time tension data is synchronized with motion attitude data collected by the inertial measurement unit (IMU) and fused at the feature layer. This fusion is then input into a pre-trained multimodal motion recognition model to identify the specific movements of the Eight-Section Brocade exercise currently being performed by the user. The multimodal motion recognition model is a neural network model trained on sample data, and its input includes tension feature vectors and attitude feature vectors. The feature layer fusion method involves extracting time-domain features (mean, variance, peak value, peak time, rise time, waveform factor) and frequency-domain features (spectral energy, dominant frequency, spectral entropy) from the tension data, and extracting attitude features (mean, variance, peak value, and zero-crossing rate of three-axis acceleration; mean, variance, peak value, and zero-crossing rate of three-axis angular velocity; pitch angle, roll angle, yaw angle, and their rate of change) from the IMU data. The two types of feature vectors are then concatenated and input into the neural network. The multimodal action recognition model uses a hybrid neural network combining 1D-CNN and LSTM. The structure includes convolutional layers, LSTM layers, fully connected layers, and an output layer (8 neurons corresponding to the eight moves of Baduanjin).

[0058] The system also includes an elastic coefficient self-calibration unit, the configuration of which has been described in detail in the first embodiment and will not be repeated here.

[0059] The third implementation method: Figure 3 A muscle strength assessment method based on the above system is shown, which includes the following steps: S0 (optional): Self-calibration of elasticity coefficient Before training begins or during training intervals, calibration tension data collected by a multi-point flexible thin-film tension sensor array is received when the user performs preset standard calibration actions. The calibration tension data is compared with the reference tension data corresponding to the pre-stored standard calibration actions to calculate the current elastic coefficient correction factor. The correction factor is applied to the correction of subsequent real-time tension data to eliminate tension data drift caused by elastic band aging or temperature changes. The specific method is the same as the first implementation method.

[0060] S1: Acquisition of raw tension signal A multi-point flexible thin-film tension sensor array integrated within the resistance elastic band is used to simultaneously acquire raw tension signals from multiple key force points during the user's Baduanjin training. Simultaneously, motion posture data is acquired synchronously through an inertial measurement unit integrated on the resistance elastic band.

[0061] S2: Signal Conditioning and Processing By integrating a signal conditioning circuit onto the resistance elastic band, the original tension signal is amplified and filtered to obtain real-time tension data.

[0062] S3: Move Recognition The system invokes pre-stored standard mechanical models for each move of the Eight-Section Brocade exercise to identify the specific move the user is currently performing based on real-time tension data. The standard mechanical model includes preset tension thresholds for the start, peak, and end times of each move, as well as a rising slope characteristic curve from the start to the peak and a falling slope characteristic curve from the peak to the end. Move identification can employ a dynamic time warping algorithm or a multimodal fusion neural network model.

[0063] If a multimodal fusion approach is adopted, real-time tension data and motion posture data are synchronized in time and fused at the feature layer. The data is then input into a pre-trained multimodal motion recognition model to obtain the classification probability of the eight moves of Baduanjin. The move with the highest probability is then taken as the recognition result.

[0064] S4: Data Segmentation Based on the temporal characteristics of the standard mechanics model corresponding to the identified specific move, the real-time tension data is automatically segmented into independent data segments corresponding to that move. The segmentation method is the same as in the second implementation method.

[0065] S5: Multi-dimensional assessment For each independent data segment, its matching degree with the corresponding standard mechanical model is calculated to generate multidimensional muscle strength assessment results. The multidimensional muscle strength assessment results include strength temporal matching degree, strength amplitude matching degree, strength duration matching degree, and bilateral synergy coefficient.

[0066] The bilateral coordination coefficient is obtained by calculating the synchronicity and amplitude ratio of the force exerted on both sides, based on time-series data from corresponding sensors on both sides acquired by a multi-point tension sensor array. The corresponding sensors on both sides refer to sensors positioned according to the symmetrical anatomical locations of the human body, such as the sensors at the left and right acromions in the first sensor group. The specific calculation method is as follows: Timing synchronization index: the signal sequence of the corresponding sensors on the left and right sides. and Perform cross-correlation analysis: Take the largest cross-correlation number .

[0067] Amplitude ratio stability index: Calculates the ratio sequence of the amplitudes of the left and right signals. Then calculate the standard deviation of the ratio sequence. and mean This leads to the coefficient of variation: ; The amplitude ratio stability index is defined as: ; Two-sided coordination coefficient: ; S6: Output Results The multi-dimensional muscle strength assessment results are output to the user through the feedback module. Output methods include visual reports (progress bars, radar charts, heatmaps), voice prompts, and historical comparisons.

[0068] In light of current practical needs, the above-described embodiments of this invention are not limited to these specific implementations. Any changes made within the scope of knowledge possessed by those skilled in the art, without departing from the concept of this invention, still fall within the protection scope of this invention.

Claims

1. A muscle strength assessment system for Baduanjin training, characterized in that, include: A resistance elastic band, comprising an elastic band body and a multi-point flexible thin-film tension sensor array integrated inside the elastic band body; The multi-point flexible thin-film tension sensor array is embedded at the key stress points of the elastic band body according to the distribution of the main force-generating muscle groups corresponding to the eight moves of Baduanjin, and according to the preset sensor layout rules. The sensor layout rules include the spacing between sensors, the layout density, and the way they are combined with the elastic band body. The resistance elastic band also includes a signal conditioning circuit, which is electrically connected to the multi-point flexible thin-film tension sensor array and integrated inside the elastic band body or at the fixed end of the elastic band body. It is used to amplify and filter the raw signals collected by the sensors. The data acquisition module is communicatively connected to the signal conditioning circuit and is used to receive and process the amplified and filtered real-time tension data. The memory stores standard mechanical models of each move of Baduanjin. The standard mechanical models are constructed based on the time-tension characteristic data collected from the standard movement samples of Baduanjin. They include the tension thresholds at the start time, peak time, and end time preset for each move of Baduanjin, as well as the rising slope characteristic curve from the start to the peak time and the falling slope characteristic curve from the peak time to the end time. When a processor executes a computer program stored in memory, it performs the following steps: Receive real-time tension data from the data acquisition module; The standard mechanical models of each move of Baduanjin stored in the memory are called to identify the move in the real-time tension data in order to determine the specific move of Baduanjin that the user is currently performing. Based on the temporal characteristics of the standard mechanical model corresponding to the identified specific move, the real-time tension data is automatically segmented into independent data segments corresponding to that move; For each independent data segment, its matching degree with the corresponding standard mechanical model is calculated to generate a multi-dimensional muscle strength assessment result, which includes the strength temporal matching degree, the strength amplitude matching degree, and the strength duration matching degree. The feedback module is used to output the multi-dimensional muscle strength assessment results in a way that is perceptible to the user.

2. The muscle strength assessment system for Baduanjin training according to claim 1, characterized in that, The sensors in the multi-point flexible thin-film tension sensor array are distributed, and the spacing between adjacent sensors is a preset spacing determined according to the concentration of the force points of the corresponding moves. The layout density is dynamically adjusted according to the concentration of the force points of the corresponding moves.

3. The muscle strength assessment system for Baduanjin training according to claim 1, characterized in that, The specific sensor layout rules are as follows: For the "Holding up the sky with both hands to regulate the three jiaos" move in Baduanjin, the first sensor group is set at the key stress points of the corresponding shoulder and back muscle groups. In response to the "drawing the bow with both hands like shooting an eagle" move, a second sensor group was installed at the key stress points of the corresponding chest, back, upper limb and lower limb muscle groups. For the "single lift to regulate the spleen and stomach" move, a third sensor group is set at the key stress points of the corresponding waist, abdomen and upper limb muscle groups. To address the "five strains and seven injuries from looking backwards" technique, a fourth sensor group is installed at key stress points corresponding to the neck, shoulder, back, and upper limb muscle groups. To counter the "shaking head and wagging tail to extinguish heart fire" move, a fifth sensor group is installed at the key stress points corresponding to the neck, waist, hips and lower limb muscle groups. For the "Two hands grasping the feet to strengthen the kidneys and waist" move, a sixth sensor group is set at the key stress points of the corresponding back and lower limb muscles. To address the "clenched fist and glaring eyes to increase strength" move, a seventh sensor group was installed at the key stress points of the corresponding upper limb, shoulder and back, and lower limb muscle groups. To address the "Seven Bounces on the Back to Cure All Diseases" technique, an eighth sensor group was installed at key stress points in the corresponding lower limbs and foot muscles.

4. The muscle strength assessment system for Baduanjin training according to claim 1, characterized in that, The processor also performs the following steps: Based on real-time tension data collected by a multi-point tension sensor array, the signal correlation between adjacent sensors is calculated. When the correlation between adjacent sensor signals is detected to be lower than a preset threshold, it is determined that a sensor position shift has occurred. In response to the judgment, a calibration prompt signal is issued, or an offset compensation algorithm is automatically executed, the offset compensation algorithm including re-establishing the mapping relationship between the sensor and the key force points.

5. The muscle strength assessment system for Baduanjin training according to claim 1, characterized in that, It also includes an inertial measurement unit, integrated on the resistive elastic band, for synchronously collecting the user's motion posture data during Baduanjin training; The processor also performs the following: time synchronization and feature layer fusion of the real-time tension data and the motion posture data, and inputs them into a pre-trained multimodal motion recognition model to identify the specific moves of Baduanjin that the user is currently performing; The multimodal motion recognition model is a neural network model trained based on sample data, and its inputs include tension feature vectors and posture feature vectors.

6. The muscle strength assessment system for Baduanjin training according to claim 1, characterized in that, It also includes an elastic coefficient self-calibration unit, which is configured as follows: Before training begins or during training intervals, the multi-point flexible thin-film tension sensor array collects calibration tension data when the user performs preset standard calibration actions. The calibration tension data is compared with the reference tension data corresponding to the standard calibration action stored in the memory, and the current elastic coefficient correction factor is calculated. The correction factor is applied to the correction of subsequent real-time tension data to eliminate tension data drift caused by elastic band aging or temperature changes.

7. The muscle strength assessment system for Baduanjin training according to claim 1, characterized in that, The results of the multidimensional muscle strength assessment include: Force timing matching degree is used to characterize the degree of consistency between the user's actual force application timing and the preset force application timing in the standard mechanical model; Force amplitude matching degree is used to characterize the percentage of the user's actual force amplitude compared to the preset force amplitude in the standard mechanical model; Strength consistency matching is used to characterize the degree to which the tension decay rate of the user during the hold phase matches the preset decay rate in the standard mechanical model.

8. A resistance elastic band for Baduanjin training, characterized in that, include: Elastic band body; The multi-point flexible thin-film tension sensor array consists of multiple flexible thin-film tension sensors arranged according to the distribution of the main force-generating muscle groups corresponding to the eight moves of Baduanjin. Each sensor is embedded in the interior of the elastic band body according to a preset sensor layout rule, which includes the spacing between sensors, the layout density, and the way they are connected to the elastic band body. The signal conditioning circuit is electrically connected to the multi-point flexible thin-film tension sensor array and is integrated inside the elastic band body or at the fixed end of the elastic band body. It is used to amplify and filter the raw signals collected by the sensors. The data interface is electrically connected to the signal conditioning circuit and is used to output the amplified and filtered real-time tension data.

9. A muscle strength assessment method based on the muscle strength assessment system for Baduanjin training according to any one of claims 1-7, characterized in that, Includes the following steps: By integrating a multi-point flexible thin-film tension sensor array inside the resistance elastic band, the original tension signals of multiple key force points are simultaneously collected when the user is performing Baduanjin training. The original tension signal is amplified and filtered by the signal conditioning circuit integrated on the resistive elastic band to obtain real-time tension data. The pre-stored standard mechanical models of each move of Baduanjin are invoked to identify the moves in the real-time tension data, so as to determine the specific move of Baduanjin that the user is currently performing. The standard mechanical model includes the tension threshold at the start time, the tension threshold at the peak time, and the tension threshold at the end time for each move of Baduanjin, as well as the rising slope characteristic curve from the start to the peak and the falling slope characteristic curve from the peak to the end. Based on the temporal characteristics of the standard mechanical model corresponding to the identified specific move, the real-time tension data is automatically segmented into independent data segments corresponding to that move; For each independent data segment, its matching degree with the corresponding standard mechanical model is calculated to generate a multi-dimensional muscle strength assessment result. The multi-dimensional muscle strength assessment result includes the strength timing matching degree, the strength amplitude matching degree, the strength duration matching degree, and the bilateral coordination coefficient. The bilateral coordination coefficient is obtained by calculating the synchronicity and amplitude ratio of the left and right sides' force exertion timing based on the time series data of the corresponding sensors on the left and right sides collected by the multi-point tension sensor array. The multidimensional muscle strength assessment results are output to the user through the feedback module.

10. The muscle strength assessment method according to claim 9, characterized in that, It also includes a self-calibration step for the elastic coefficient: Before training begins or during training intervals, the multi-point flexible thin-film tension sensor array collects calibration tension data when the user performs preset standard calibration actions. The calibration tension data is compared with the reference tension data corresponding to the pre-stored standard calibration action to calculate the current elastic coefficient correction factor; The correction factor is applied to the correction of subsequent real-time tension data to eliminate tension data drift caused by elastic band aging or temperature changes.

Images