Data-based joint motion monitoring methods, devices, and storage media

Abstract

This invention relates to the field of data processing technology, and more particularly to a method, device, and storage medium for joint motion monitoring based on data analysis. The method includes: identifying gait symmetry features based on hip and knee joint angles, and extracting joint angle features within a gait window; extracting muscle fatigue features based on surface electromyography (EMG) signals; determining device interaction anomaly features based on joint drive currents; determining a comprehensive motion posture index based on gait symmetry and gait stability features; determining a comprehensive exercise load index based on muscle fatigue and device interaction anomaly features, and combining the comprehensive motion posture index to determine joint motion health, and issuing an early warning to the user. This invention integrates multimodal sensor data to comprehensively assess joint motion health from three dimensions: posture, fatigue, and interaction, and achieves intelligent early warning.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, device, and storage medium for joint motion monitoring based on data analysis. Background Technology

[0002] Exoskeleton robots and rehabilitation training devices are playing an increasingly important role in assisting walking, gait training, and motor function rehabilitation. However, existing monitoring technologies mostly rely on single sensor data (such as detecting only joint angles or plantar pressure), making it difficult to comprehensively assess the complex interaction between the human body and the exoskeleton. Relying solely on kinematic parameters cannot detect muscle fatigue accumulation in a timely manner, while electromyographic signals alone are easily affected by electrode placement and skin impedance. Furthermore, improper device wearing or mechanical jamming often manifests as abnormal fluctuations in the drive current, a dimension that current solutions lack effective utilization of.

[0003] More importantly, existing technologies lack an early warning mechanism that can integrate multi-source information, distinguish between normal and abnormal movement patterns, and provide a comprehensive health assessment. This results in a high false alarm rate, poor user experience, and an inability to meet the reliability requirements for clinical rehabilitation and daily assistance. Summary of the Invention

[0004] The purpose of this invention is to provide a joint motion monitoring method, device, and storage medium based on data analysis, so as to solve at least one of the problems existing in the prior art.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A data analysis-based method for monitoring joint motion includes:

[0007] Gait symmetry features are identified based on hip and knee joint angles, and joint angle features within the gait window are extracted.

[0008] Muscle fatigue characteristics are extracted based on surface electromyography signals;

[0009] Determine abnormal device interaction characteristics based on joint drive current;

[0010] The comprehensive index of movement posture is determined based on gait symmetry and gait stability characteristics;

[0011] The comprehensive exercise load index is determined based on muscle fatigue characteristics and abnormal device interaction characteristics, and the joint health status is determined by combining the comprehensive exercise posture index, and an early warning is issued to the user.

[0012] Furthermore, within each gait window, the time series of hip joint angle θh, knee joint angle θk, forefoot pressure Ff, and heel pressure Fh were analyzed:

[0013] Threshold cross-detection is performed on the heel pressure signal Fh and the forefoot pressure signal Ff to determine the heel strike event and the toe lift event. Using these events, combined with the local peak value of the hip joint angle θh, the left gait cycle and the right gait cycle within the window are identified.

[0014] Furthermore, for the i-th identified left gait cycle, record its cycle duration Tli; for the i-th right gait cycle, record its cycle duration Tri; calculate the average duration Tla of all left gait cycles within the gait window, and the average duration Tra of all right gait cycles, then the gait symmetry feature Gs = Tla / Tra.

[0015] Based on the changes in plantar pressure, each gait cycle is divided into a support phase and a swing phase. The position of the plantar pressure center in the anterior-posterior direction at each sampling moment is calculated as follows: CP = (Ff × Lf + Fh × Lh) / (Ff + Fh), where Lf and Lh are the distances from the center point of the forefoot and heel pressure sensors to the heel reference point, respectively. The standard deviation σCP of CP during the support phase within this gait window is extracted. σCP is divided by the wearer's leg length Lg to obtain the gait stability feature St.

[0016] Furthermore, within each real-time cycle, the acquired rectus femoris electromyography (sEMGr) and gastrocnemius muscle electromyography (sEMGg) signals are processed:

[0017] The raw electromyography (EMG) signal is bandpass filtered, and the filtered EMG signal is subjected to short-time Fourier transform. The median frequency MFr of the rectus femoris EMG signal and the median frequency MFg of the gastrocnemius EMG signal within the current real-time period are calculated, thereby determining the first muscle fatigue factor Fr and the second muscle fatigue factor Fg.

[0018] Fr=min(1,max(0,(MFrb-MFr) / MFrb));

[0019] Fg=min(1,max(0,(MFgb-MFg) / MFgb));

[0020] The muscle fatigue characteristic Ft is determined based on the first muscle fatigue factor Fr and the second muscle fatigue factor Fg, where Ft = max(Fr, Fg).

[0021] Wherein, MFrb is the first preset muscle fatigue baseline, and MFgb is the second preset muscle fatigue baseline.

[0022] Furthermore, the hip joint drive current Ih and knee joint drive current Ik are analyzed within each real-time cycle:

[0023] Calculate the average value of the standard deviation σIh of the hip joint drive current within the real-time cycle, and similarly calculate the average value of the standard deviation σIk of the knee joint drive current, and then determine the abnormal interaction characteristics Ca of the equipment, Ca=max{σIh / Iha,σIk / Ika}.

[0024] Furthermore, at the end of each gait window, the comprehensive motion posture index Pi is determined based on the gait symmetry feature Gs and the gait stability feature St acquired within the window.

[0025] The gait symmetry score Ps is determined based on the gait symmetry feature Gs, where Ps = 1 - min(1, |Gs-1|).

[0026] The stability score Pst is determined based on the gait stability feature St, where Pst = 1 - min(1, St / sth), and sth is a preset stability threshold.

[0027] The comprehensive index of motion posture, Pi, is determined based on the gait symmetry score Ps and the stability score Pst, where Pi = w1 × Ps + w2 × Pst.

[0028] Where w1 is the gait symmetry weight, w2 is the stability weight, and w1+w2=1.

[0029] Furthermore, at the end of each gait window, the arithmetic mean Ftavg of the comprehensive muscle fatigue score Ft for all real-time cycles within that gait window, and the arithmetic mean Caavg of the device interaction anomaly features Ca for all real-time cycles are calculated, and the device interaction anomaly score Cs is determined, thereby determining the comprehensive exercise load index Li.

[0030] Cs=lg[5×min(1,Caavg / rc)+1] / lg6;

[0031] Li = α1 × Ftavg + α2 × Cs;

[0032] Where α1 is the fatigue weight, α2 is the equipment anomaly weight, and α1+α2=1, and rc is the preset interaction feature threshold.

[0033] Furthermore, the comprehensive index of movement posture and the comprehensive index of exercise load are integrated to determine the joint movement health H, H=β1×Pi+β2×(1-Li);

[0034] Where β1 is the attitude weight, β2 is the load weight, and β1+β2=1;

[0035] The joint movement health status H is compared with the preset warning threshold ha. If H is greater than or equal to ha, no warning is issued to the user; otherwise, a movement mode warning is issued to the user.

[0036] According to another aspect of this application, a joint motion monitoring device based on data analysis is provided, comprising:

[0037] Gait symmetry unit is used to identify gait symmetry features based on hip and knee joint angles and extract joint angle features within the gait window;

[0038] The electromyography fatigue unit is used to extract muscle fatigue characteristics based on surface electromyography signals.

[0039] An interaction anomaly unit is used to determine the characteristics of device interaction anomalies based on the joint drive current.

[0040] The posture synthesis unit is used to determine the comprehensive index of motion posture based on gait symmetry characteristics and gait stability characteristics;

[0041] The health warning unit is used to determine the comprehensive exercise load index based on muscle fatigue characteristics and abnormal device interaction characteristics, and to determine the joint health status by combining the comprehensive exercise posture index, and to issue a warning to the user.

[0042] According to another aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, wherein the computer program is used to control the electronic device in which the computer-readable storage medium is located to perform the data analysis-based joint motion monitoring method during runtime.

[0043] The beneficial effects of this invention are as follows: By integrating four types of sensor data—joint angle, surface electromyography, plantar pressure, and driving current—a multi-level evaluation system is constructed from four dimensions: gait symmetry, gait stability, muscle fatigue, and device interaction anomalies. The method employs a dual-timescale architecture of real-time cycle and gait window, balancing response speed and evaluation stability. Automatic segmentation and feature quantification of gait cycles are achieved through gait event detection and feature extraction; non-invasive real-time monitoring of muscle fatigue is realized through median frequency analysis; and rapid identification of device interaction anomalies is achieved through current coefficient of variation analysis. Finally, the data is fused into joint movement health status and triggers graded early warnings, effectively solving the problems of one-sided single-parameter monitoring, high false alarm rate, and lack of comprehensive evaluation in existing technologies. This provides a reliable means of motion safety monitoring for exoskeleton robots and rehabilitation training equipment. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a flowchart illustrating the joint motion monitoring method based on data analysis in this embodiment.

[0046] Figure 2 This is a flowchart illustrating the exercise health analysis method in this embodiment.

[0047] Figure 3 This is a schematic diagram of the joint motion monitoring device based on data analysis in this embodiment. Detailed Implementation

[0048] To more clearly illustrate the present invention, the following description, in conjunction with preferred embodiments and accompanying drawings, further explains the invention. Similar components in the drawings are indicated by the same reference numerals. Those skilled in the art should understand that the specific description below is illustrative rather than restrictive and should not be construed as limiting the scope of protection of the present invention.

[0049] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0050] Specifically, this embodiment is applied to exoskeleton robots or rehabilitation training equipment to monitor the movement status of the human hip and knee joints in real time and to provide early warnings of potential abnormal movement patterns, excessive fatigue, and improper device wearing.

[0051] Please see Figure 1 As shown, this is a flowchart illustrating the joint motion monitoring method based on data analysis in this embodiment. Before the method is executed, the system synchronously collects data through a sensor array deployed on the exoskeleton joints and the wearer's body surface, including:

[0052] Before data acquisition begins, the magnetic absolute angle encoder needs to undergo initial zero-point calibration. Specifically, the wearer stands upright on a horizontal surface, with both the hip and knee joints at their physiological zero positions (hip 0°, knee 0°). A "zero" command is sent through the exoskeleton control interface, recording the current encoder reading as the system zero point. Subsequent θh and θk measurements are corrected for offset based on this zero point. If a fixed deviation occurs between the encoder's mechanical zero point and the human physiological zero point due to device disassembly, assembly, or improper wearing, the system will automatically execute the above calibration process upon each power-on initialization, or after manual triggering by the user. The data includes:

[0053] Joint angle data: The hip joint angle value θh and the knee joint angle value θk are collected in real time by magnetic absolute angle encoders installed at the hip and knee joint pivots to form a time series.

[0054] Surface electromyography data: Raw time-series data of electromyography signals were acquired by attaching dry electrode pads to the belly of the rectus femoris muscle in the thigh and the gastrocnemius muscle in the calf.

[0055] Plantar pressure data: The forefoot pressure value Ff and the heel pressure value Fh are collected by thin-film pressure sensors built into the forefoot and heel areas of the insole;

[0056] Device interaction data: Read the effective value of hip joint drive current Ih and knee joint drive current Ik from the control circuit of the exoskeleton joint actuator;

[0057] This embodiment does not specifically limit the communication protocol and transmission method of the above data; those skilled in the art can freely set them according to the on-site conditions. The method uses 1 second as a real-time analysis cycle and 30 seconds as a gait cycle analysis window for cyclical calculation and monitoring. For ease of distinction, the "real-time analysis cycle" will be abbreviated as "real-time cycle" and the "gait cycle analysis window" will be abbreviated as "gait window" below.

[0058] The method includes:

[0059] Step S1: Identify gait symmetry features based on hip and knee joint angles, and extract joint angle features within the gait window.

[0060] Specifically, within each gait window, the time series of hip joint angle θh, knee joint angle θk, forefoot pressure Ff, and heel pressure Fh are analyzed:

[0061] Threshold cross-detection is performed on the heel pressure signal Fh and the forefoot pressure signal Ff to determine heel strike events and toe-off events.

[0062] If the heel pressure Fh(t) at the current sampling time t is greater than or equal to the contact threshold Fth, and the heel pressure Fh(t-Δt) at the previous sampling time t-Δt is less than Fth, that is, when the pressure value rises from less than the threshold to greater than or equal to the threshold and the rising edge crosses zero, it is determined to be a heel strike event.

[0063] If the plantar forefoot pressure Ff(t) at the current sampling time t is less than or equal to the contact threshold Fth, and the plantar forefoot pressure Ff(t-Δt) at the previous sampling time t-Δt is greater than or equal to Fth, then the event is determined as toe-off event when the pressure value drops from greater than the threshold to less than or equal to the threshold and the falling edge crosses zero.

[0064] Using the above events, combined with the local peak value of the hip joint angle θh, the left and right gait cycles within this window are identified:

[0065] Left gait cycle identification: Taking the left heel strike event as the cycle start point, the search for the first subsequent left toe lift event as the end of the support phase, and then the search for the first subsequent left heel strike event as the start point of the next cycle. The time interval between two adjacent left heel strike events is the duration Tli of a left gait cycle. To eliminate false events, at least one left toe lift event and a local peak of the hip joint angle θh must be detected between two left heel strike events. Otherwise, it is considered an invalid cycle and discarded. The local peak of the hip joint angle θh is obtained by using a sliding window peak finding algorithm. The peak is defined as the point with the largest amplitude among the five sampling points before and after, and the peak value is greater than the minimum angle threshold of the swing phase of 20°.

[0066] Right gait cycle identification: The same logic is executed synchronously on the right plantar pressure signal to obtain the right gait cycle duration Tri;

[0067] Each gait cycle is defined as the process from the left heel strike to the next left heel strike, while simultaneously identifying the right gait cycle;

[0068] For the i-th identified left gait cycle, record its cycle duration Tli; for the i-th right gait cycle, record its cycle duration Tri; calculate the average duration Tla of all left gait cycles within the gait window, and the average duration Tra of all right gait cycles, then the gait symmetry feature Gs = Tla / Tra.

[0069] Simultaneously, based on changes in plantar pressure, each gait cycle is divided into a support phase (from heel strike to toe lift) and a swing phase (from toe lift to the next heel strike). The position of the plantar pressure center in the anterior-posterior direction at each sampling moment is calculated as follows: CP = (Ff × Lf + Fh × Lh) / (Ff + Fh), where Lf and Lh are the distances from the center point of the forefoot and heel pressure sensors to the heel reference point, respectively. The standard deviation σCP of CP during the support phase within the gait window is extracted. σCP is divided by the wearer's leg length Lg to obtain the gait stability characteristic St. The support phase is from heel strike to toe lift, and the swing phase is from toe lift to the next heel strike.

[0070] Preferably, the threshold Fth is 5N, the sampling period Δt is 0.01s, the sliding window length for peak detection is set to 11 sampling points, and the minimum peak height is set to 20° to exclude false peaks during rest or slight swaying. If the peak height is less than 20° due to individual differences or special gait, the system allows the user to lower the threshold to 10° during initialization.

[0071] Preferably, the leg length Lg is manually input by the wearer through the host computer interface (in meters), or is obtained by automatically measuring the total extension length of the thigh and calf linkage by the laser rangefinder on the exoskeleton mechanical structure; the distances Lf and Lh from the center point of the pressure sensor to the heel reference point are determined by the insole size: in this embodiment, Lf=0.18m and Lh=0.05m are suitable for size 38 insoles; if other sizes are used, they are scaled linearly proportionally.

[0072] Specifically, by fusing hip and knee joint angles with plantar pressure signals, high-precision segmentation of the gait cycle and quantitative extraction of gait features were achieved. Threshold cross-detection was used to accurately identify heel strike and toe-off events, and the peak hip joint angle was combined to verify cycle integrity, effectively eliminating spurious event interference and ensuring the reliability of gait cycle segmentation. Gait symmetry features were obtained based on the ratio of left and right gait cycle durations, objectively reflecting the coordination of the wearer's bilateral lower limb movements. Gait stability features were obtained by using the ratio of the standard deviation of the plantar pressure center in the stance phase to leg length, eliminating the influence of individual body size differences on stability assessment. This step provides fundamental feature input for subsequent comprehensive evaluation of movement posture, enabling the exoskeleton robot to perceive changes in the quality of human gait in real time.

[0073] Please continue reading. Figure 1 As shown, the data analysis-based joint motion monitoring method further includes:

[0074] Step S2: Extract muscle fatigue characteristics based on surface electromyography signals.

[0075] Specifically, within each real-time cycle, the acquired rectus femoris electromyography (sEMGr) and gastrocnemius muscle electromyography (sEMGg) signals are processed:

[0076] The raw electromyography (EMG) signals were bandpass filtered (20-450Hz), and the filtered EMG signals were subjected to short-time Fourier transform. The median frequency MFr of the rectus femoris EMG signal and the median frequency MFg of the gastrocnemius EMG signal within the current real-time period were calculated, thereby determining the first muscle fatigue factor Fr and the second muscle fatigue factor Fg.

[0077] Fr=min(1,max(0,(MFrb-MFr) / MFrb));

[0078] Fg=min(1,max(0,(MFgb-MFg) / MFgb));

[0079] The muscle fatigue characteristic Ft is determined based on the first muscle fatigue factor Fr and the second muscle fatigue factor Fg, where Ft = max(Fr, Fg).

[0080] Wherein, MFrb is the first preset muscle fatigue baseline, and MFgb is the second preset muscle fatigue baseline.

[0081] Preferably, the short-time Fourier transform uses a Hanning window with a window length of 256 sampling points, an overlap rate of 50%, and a frequency resolution of approximately 3.9 Hz, based on which the median frequency within each window is calculated.

[0082] Preferably, the first preset muscle fatigue baseline is the typical median frequency value of the rectus femoris muscle during moderate-intensity contraction in a healthy adult. It is recommended that this value be specified as the initial value measured during device initialization, when the user is standing at rest, or when swinging their leg without load, or that the empirical value of 80Hz be used directly. The second preset muscle fatigue baseline is the typical median frequency value of the gastrocnemius muscle during moderate-intensity contraction in a healthy adult. It is recommended that this value be specified as the initial value measured during device initialization, when the user is standing at rest, or when swinging their leg without load, or that the empirical value of 85Hz be used directly.

[0083] Specifically, by performing short-time Fourier transform on the surface electromyography (EMG) signals of the rectus femoris and gastrocnemius muscles, the median frequency, a feature parameter highly correlated with muscle fatigue, was extracted. Compared to time-domain amplitude features, median frequency is more sensitive to metabolic muscle fatigue and less susceptible to interference from skin impedance and electrode position drift. By comparing the current median frequency with a preset fatigue baseline and performing amplitude limiting, a normalized muscle fatigue factor was obtained, eliminating individual differences in muscle characteristics. Taking the maximum value of the two muscle fatigue factors as the comprehensive fatigue feature can reflect the state of the most fatigued muscle groups and avoid the averaging effect masking local over-fatigue. This step provides a direct physiological indicator for exercise load assessment and helps prevent compensatory exercise injuries caused by muscle fatigue.

[0084] Please continue reading. Figure 1 As shown, the data analysis-based joint motion monitoring method further includes:

[0085] Step S3: Determine the abnormal characteristics of device interaction based on the joint drive current.

[0086] Specifically, the hip joint drive current Ih and the knee joint drive current Ik are analyzed in each real-time cycle:

[0087] Calculate the average value of the standard deviation σIh of the hip joint drive current within the real-time cycle, and similarly calculate the average value of the standard deviation σIk of the knee joint drive current, and then determine the abnormal interaction characteristics Ca of the equipment, Ca=max{σIh / Iha,σIk / Ika}.

[0088] Specifically, the ratio of the standard deviation to the mean of the exoskeleton joint drive current is used as a feature of abnormal device interaction. The coefficient of variation construction makes this feature insensitive to load changes, highlighting the relative severity of current fluctuations, thus effectively identifying abnormal torque fluctuations caused by mechanical jamming, sensor malfunction, or loose fitting. This feature is simple to calculate, highly real-time, and can be obtained from existing control loops without additional hardware, demonstrating good engineering applicability. By monitoring this feature, the system can respond promptly to abnormalities in the interaction force between the device and the human body, providing a mechanical basis for subsequent motion load assessment and early warning.

[0089] Please continue reading. Figure 1 As shown, the data analysis-based joint motion monitoring method further includes:

[0090] Step S4: Determine the comprehensive index of movement posture based on gait symmetry characteristics and gait stability characteristics.

[0091] Specifically, at the end of each gait window, the comprehensive motion posture index Pi is determined based on the gait symmetry feature Gs and the gait stability feature St acquired within the window.

[0092] The gait symmetry score Ps is determined based on the gait symmetry feature Gs, where Ps = 1 - min(1, |Gs-1|).

[0093] The stability score Pst is determined based on the gait stability feature St, where Pst = 1 - min(1, St / sth), and sth is a preset stability threshold.

[0094] The comprehensive index of motion posture, Pi, is determined based on the gait symmetry score Ps and the stability score Pst, where Pi = w1 × Ps + w2 × Pst.

[0095] Where w1 is the gait symmetry weight, w2 is the stability weight, and w1+w2=1.

[0096] Preferably, in this embodiment, the gait symmetry weight is 0.6, the stability weight is 0.4, and the preset stability threshold is 0.06.

[0097] Specifically, gait symmetry and gait stability features are integrated into a single comprehensive movement posture index. Through a weighted combination of symmetry and stability scores, this index can comprehensively quantify the wearer's movement posture quality from two dimensions: spatiotemporal symmetry and center of pressure stability. This provides an intuitive and comprehensive quantitative indicator for evaluating the movement patterns of exoskeletons, facilitating subsequent health calculations and early warning decisions.

[0098] Please continue reading. Figure 1 As shown, the data analysis-based joint motion monitoring method further includes:

[0099] Step S5: Determine the comprehensive exercise load index based on muscle fatigue characteristics and abnormal device interaction characteristics, and combine it with the comprehensive exercise posture index to determine joint movement health, and issue a warning to the user.

[0100] Please see Figure 2 As shown, the method for analyzing exercise health includes:

[0101] Step S51: Determine the comprehensive exercise load index based on muscle fatigue characteristics and abnormal equipment interaction characteristics.

[0102] Specifically, at the end of each gait window, the arithmetic mean Ftavg of the comprehensive muscle fatigue score Ft for all real-time cycles within that gait window, and the arithmetic mean Caavg of the device interaction anomaly features Ca for all real-time cycles are calculated, and the device interaction anomaly score Cs is determined, thereby determining the comprehensive exercise load index Li.

[0103] Cs=lg[5×min(1,Caavg / rc)+1] / lg6;

[0104] Li = α1 × Ftavg + α2 × Cs;

[0105] Where α1 is the fatigue weight, α2 is the equipment anomaly weight, and α1+α2=1, and rc is the preset interaction feature threshold.

[0106] Preferably, in this embodiment, the fatigue weight is 0.7, the equipment abnormality weight is 0.3, and the preset interaction feature threshold is 0.3.

[0107] Specifically, muscle fatigue characteristics and abnormal device interaction characteristics are fused into a comprehensive exercise load index. Muscle fatigue characteristics reflect the body's own physiological load level, while abnormal device interaction characteristics reflect the mechanical interaction load level between the exoskeleton and the body. Both characterize the burden on the musculoskeletal system from physiological and mechanical perspectives, respectively. Through weighted fusion, this index comprehensively reflects the overall pressure of current exercise on the wearer's joints and muscular systems. Using an arithmetic average to aggregate real-time data within the gait window effectively suppresses the influence of instantaneous noise on the evaluation results, making the load index more stable and reliable.

[0108] Please continue reading. Figure 2 As shown, the exercise health analysis method further includes:

[0109] Step S52: Determine the joint health status based on the comprehensive index of exercise load and the comprehensive index of exercise posture, and issue a warning to the user.

[0110] Specifically, the comprehensive index of movement posture and the comprehensive index of exercise load are integrated to determine the joint movement health H, H=β1×Pi+β2×(1-Li);

[0111] Where β1 is the attitude weight, β2 is the load weight, and β1+β2=1;

[0112] The joint movement health status H is compared with the preset warning threshold ha. If H is greater than or equal to ha, no warning is issued to the user; otherwise, a movement mode warning is issued to the user.

[0113] Preferably, in this embodiment, the attitude weight is 0.8, the load weight is 0.2, and the preset warning threshold is 0.65.

[0114] Specifically, the comprehensive index of movement posture and the comprehensive index of exercise load are fused together to obtain the joint movement health score. This health score considers both the quality of movement posture and the magnitude of exercise load; a higher posture index indicates better health, and a lower load index indicates better health, achieving a two-way evaluation through complementary weighting. The health score is compared with a preset warning threshold, and a warning is triggered only when the health score falls below the threshold, avoiding frequent false alarms. This warning mechanism can distinguish between postural fluctuations during normal training and abnormal states that truly require intervention, providing users with timely and accurate feedback on movement patterns and effectively reducing the risk of secondary injuries caused by over-fatigue and improper wear.

[0115] Please see Figure 3 As shown, the data analysis-based joint motion monitoring device includes:

[0116] Gait symmetry unit is used to identify gait symmetry features based on hip and knee joint angles and extract joint angle features within the gait window;

[0117] The electromyography fatigue unit is used to extract muscle fatigue characteristics based on surface electromyography signals.

[0118] An interaction anomaly unit is used to determine the characteristics of device interaction anomalies based on the joint drive current.

[0119] The posture synthesis unit is used to determine the comprehensive index of motion posture based on gait symmetry characteristics and gait stability characteristics;

[0120] The health warning unit is used to determine the comprehensive exercise load index based on muscle fatigue characteristics and abnormal device interaction characteristics, and to determine the joint health status by combining the comprehensive exercise posture index, and to issue a warning to the user.

[0121] The data analysis-based joint motion monitoring device provided in this application can execute the data analysis-based joint motion monitoring method provided in any embodiment of this application, and has the corresponding functional modules and beneficial effects of the execution method.

[0122] This application also provides a computer-readable storage medium, which is a tangible physical storage medium that can store the aforementioned computer program and various types of data used in the program; the physical storage medium includes, but is not limited to, existing physical storage media or combinations thereof, such as random access memory, read-only memory, optical disk, and hard disk.

[0123] Those skilled in the art will understand that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technique for storing information (such as computer-readable programs, data structures, program modules, or other data). Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable programs, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and can include any information delivery medium.

[0124] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. For those skilled in the art, other variations or modifications can be made based on the above description. It is impossible to exhaustively list all the implementation methods here. All obvious variations or modifications derived from the technical solutions of the present invention are still within the protection scope of the present invention.

Claims

1. A joint motion monitoring method based on data analysis, characterized in that, include: Gait symmetry features are identified based on hip and knee joint angles, and joint angle features within the gait window are extracted. Muscle fatigue characteristics are extracted based on surface electromyography signals; Determine abnormal device interaction characteristics based on joint drive current; The comprehensive index of movement posture is determined based on gait symmetry and gait stability characteristics; The comprehensive exercise load index is determined based on muscle fatigue characteristics and abnormal device interaction characteristics, and the joint health status is determined by combining the comprehensive exercise posture index, and an early warning is issued to the user.

2. The joint motion monitoring method based on data analysis according to claim 1, characterized in that, Within each gait window, the time series of hip angle θh, knee angle θk, forefoot pressure Ff, and heel pressure Fh were analyzed: Threshold cross-detection is performed on the heel pressure signal Fh and the forefoot pressure signal Ff to determine the heel strike event and the toe lift event. Using these events, combined with the local peak value of the hip joint angle θh, the left gait cycle and the right gait cycle within the window are identified.

3. The joint motion monitoring method based on data analysis according to claim 2, characterized in that, For the i-th identified left gait cycle, record its cycle duration Tli; for the i-th right gait cycle, record its cycle duration Tri; calculate the average duration Tla of all left gait cycles within the gait window, and the average duration Tra of all right gait cycles, then the gait symmetry feature Gs = Tla / Tra. Based on the changes in plantar pressure, each gait cycle is divided into a support phase and a swing phase. The position of the plantar pressure center in the anterior-posterior direction at each sampling moment is calculated as follows: CP = (Ff × Lf + Fh × Lh) / (Ff + Fh), where Lf and Lh are the distances from the center point of the forefoot and heel pressure sensors to the heel reference point, respectively. The standard deviation σCP of CP during the support phase within this gait window is extracted. σCP is divided by the wearer's leg length Lg to obtain the gait stability feature St.

4. The joint motion monitoring method based on data analysis according to claim 3, characterized in that, Within each real-time cycle, the acquired rectus femoris electromyography (sEMGr) and gastrocnemius muscle electromyography (sEMGg) signals are processed: The raw electromyography (EMG) signal is bandpass filtered, and the filtered EMG signal is subjected to short-time Fourier transform. The median frequency MFr of the rectus femoris EMG signal and the median frequency MFg of the gastrocnemius EMG signal within the current real-time period are calculated, thereby determining the first muscle fatigue factor Fr and the second muscle fatigue factor Fg. Fr=min(1,max(0,(MFrb-MFr) / MFrb)); Fg=min(1,max(0,(MFgb-MFg) / MFgb)); The muscle fatigue characteristic Ft is determined based on the first muscle fatigue factor Fr and the second muscle fatigue factor Fg, where Ft = max(Fr, Fg). Wherein, MFrb is the first preset muscle fatigue baseline, and MFgb is the second preset muscle fatigue baseline.

5. The joint motion monitoring method based on data analysis according to claim 4, characterized in that, Within each real-time cycle, the hip joint drive current Ih and the knee joint drive current Ik are analyzed: Calculate the average value of the standard deviation σIh of the hip joint drive current within the real-time cycle, and similarly calculate the average value of the standard deviation σIk of the knee joint drive current, and then determine the abnormal interaction characteristics Ca of the equipment, Ca=max{σIh / Iha,σIk / Ika}.

6. The joint motion monitoring method based on data analysis according to claim 5, characterized in that, At the end of each gait window, the motion posture comprehensive index Pi is determined based on the gait symmetry feature Gs and the gait stability feature St acquired within the window. The gait symmetry score Ps is determined based on the gait symmetry feature Gs, where Ps = 1 - min(1, |Gs-1|). The stability score Pst is determined based on the gait stability feature St, where Pst = 1 - min(1, St / sth), and sth is a preset stability threshold. The comprehensive index of motion posture, Pi, is determined based on the gait symmetry score Ps and the stability score Pst, where Pi = w1 × Ps + w2 × Pst. Where w1 is the gait symmetry weight, w2 is the stability weight, and w1+w2=1.

7. The joint motion monitoring method based on data analysis according to claim 6, characterized in that, At the end of each gait window, the arithmetic mean Ftavg of the comprehensive muscle fatigue score Ft for all real-time cycles within that gait window, and the arithmetic mean Caavg of the device interaction anomaly features Ca for all real-time cycles are calculated. The device interaction anomaly score Cs is then determined, and the comprehensive exercise load index Li is determined accordingly. Cs=lg[5×min(1,Caavg / rc)+1] / lg6; Li = α1 × Ftavg + α2 × Cs; Where α1 is the fatigue weight, α2 is the equipment anomaly weight, and α1+α2=1, and rc is the preset interaction feature threshold.

8. The joint motion monitoring method based on data analysis according to claim 7, characterized in that, The joint health status H is determined by integrating the comprehensive index of movement posture and the comprehensive index of exercise load, where H = β1 × Pi + β2 × (1 - Li). Where β1 is the attitude weight, β2 is the load weight, and β1+β2=1; The joint movement health status H is compared with the preset warning threshold ha. If H is greater than or equal to ha, no warning is issued to the user; otherwise, a movement mode warning is issued to the user.

9. A joint motion monitoring device based on data analysis, applied to the joint motion monitoring method based on data analysis as described in any one of claims 1-8, characterized in that, include: Gait symmetry unit is used to identify gait symmetry features based on hip and knee joint angles and extract joint angle features within the gait window; The electromyography fatigue unit is used to extract muscle fatigue characteristics based on surface electromyography signals. An interaction anomaly unit is used to determine the characteristics of device interaction anomalies based on the joint drive current. The posture synthesis unit is used to determine the comprehensive index of motion posture based on gait symmetry characteristics and gait stability characteristics; The health warning unit is used to determine the comprehensive exercise load index based on muscle fatigue characteristics and abnormal device interaction characteristics, and to determine the joint health status by combining the comprehensive exercise posture index, and to issue a warning to the user.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein the computer program is used to control the electronic device on which the computer-readable storage medium is located to perform the joint motion monitoring method based on data analysis as described in any one of claims 1-8 during runtime.

Images