Motion state monitoring method, device and system based on high-precision sensing

Abstract

This invention relates to the field of data processing technology, specifically to a method, device, and system for motion state monitoring based on high-precision sensing. It solves the technical problem of low accuracy in assessing exercise fatigue states in existing technologies. The method includes: acquiring a user's physiological signals, a first motion signal, and a second motion signal; determining the user's real-time physiological fatigue index based on the physiological signals; determining a wear-based reliability weight for assessing the reliability of the physiological signals based on the first and second motion signals; determining a historical estimated fatigue index based on the second motion signal and historical motion data; and determining the user's fused fatigue index by weighted summation of the real-time physiological fatigue index and the historical estimated fatigue index based on the wear-based reliability weight. This invention is used for the identification and assessment of exercise fatigue states.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically to a method, device, and system for motion state monitoring based on high-precision sensing. Background Technology

[0002] With the increasing awareness of fitness among the general public and the development of smart wearable devices, health monitoring technology during exercise has received growing attention. Modern fitness equipment is transforming towards intelligent operation, and adjusting exercise intensity, reminding users to rest, and avoiding health risks from excessive exercise by monitoring their exercise status has become an important research objective. Currently, common methods for monitoring exercise status mainly rely on contact sensors to collect users' physiological signals, such as heart rate sensors and electromyography (EMG) sensors to obtain data on heart rate variability and muscle electrical activity, thereby assessing the user's level of exercise fatigue. However, in actual exercise scenarios, especially during high-intensity or high-amplitude exercise, the contact state between the sensor and the skin surface can easily change, leading to a decrease in the quality of the collected physiological signals. This results in low accuracy of current technologies in assessing exercise fatigue. Summary of the Invention

[0003] To address the technical problem that changes in the contact state between the sensor and the skin surface can degrade the quality of collected physiological signals, leading to low accuracy in assessing exercise fatigue in existing technologies, this invention aims to provide a method, device, and system for motion state monitoring based on high-precision sensing. The specific technical solution adopted is as follows: In a first aspect, the present invention provides a motion state monitoring method based on high-precision sensing. The method includes: acquiring a user's physiological signals, a first motion signal, and a second motion signal; the physiological signals characterizing the user's muscle fatigue level and autonomic nervous system fatigue load; the first motion signal characterizing the user's motion state; the second motion signal characterizing the device's motion state; determining the user's real-time physiological fatigue index based on the physiological signals; the real-time physiological fatigue index characterizing the user's current fatigue level based on physiological responses; determining a wear-based reliability weight for assessing the reliability of the physiological signals based on the first and second motion signals; determining a historical estimated fatigue index based on the second motion signal and historical motion data; the historical estimated fatigue index characterizing the user's fatigue level determined based on historical motion data; and weighted summing the real-time physiological fatigue index and the historical estimated fatigue index based on the wear-based reliability weight to determine the user's fused fatigue index.

[0004] In conjunction with the first aspect mentioned above, in one possible implementation, the physiological signals include: electromyography (EMG) signals and heart rate signals; the method specifically includes: determining an EMG fatigue factor based on the integral value of the EMG signal within a preset time window relative to the EMG integral reference value in the resting state; determining a heart rate fatigue factor based on a first attenuation factor and a second attenuation factor of the heart rate signal within the preset time window; the first attenuation factor being the ratio of the high-frequency component integral value of the heart rate signal within the preset time window to the high-frequency integral reference value in the resting state; the second attenuation factor being the ratio of the low-frequency component integral value of the heart rate signal within the preset time window to the low-frequency integral reference value in the resting state; and determining a real-time physiological fatigue index based on the EMG fatigue factor and the heart rate fatigue factor.

[0005] In conjunction with the first aspect mentioned above, in one possible implementation, the method specifically includes: determining the motion intensity of the user terminal based on a first motion signal; determining the motion intensity of the device terminal based on a second motion signal; and determining the wearing confidence weight based on the motion intensity of the user terminal and the motion intensity of the device terminal.

[0006] In conjunction with the first aspect mentioned above, in one possible implementation, the method specifically includes: when the wearing confidence weight is greater than or equal to a first confidence threshold, saving the real-time physiological fatigue index and the corresponding second motion signal as a set of historical data for constructing or updating historical motion data; when the wearing confidence weight is less than the first confidence threshold but greater than a second confidence threshold, retrieving a historical second motion signal from the historical motion data that has a similarity to the current second motion signal that meets a preset condition based on the current second motion signal; the first confidence threshold is greater than the second confidence threshold; and determining the historical real-time physiological fatigue index corresponding to the retrieved historical second motion signal as the historical estimated fatigue index.

[0007] In conjunction with the first aspect mentioned above, in one possible implementation, the method specifically includes: determining the similarity between the current second motion signal and each historical second motion signal in the historical motion data according to a sequence similarity analysis algorithm; and determining the second motion signal with the highest similarity as the historical second motion signal.

[0008] In conjunction with the first aspect mentioned above, in one possible implementation, the method specifically includes: determining a first weight of the real-time physiological fatigue index and a second weight of the historical estimated fatigue index based on the wear confidence weight; the first weight is positively correlated with the wear confidence weight; the second weight is negatively correlated with the wear confidence weight; and determining the fused fatigue index based on the real-time physiological fatigue index and its first weight, the historical estimated fatigue index and its second weight.

[0009] In conjunction with the first aspect above, in one possible implementation, the physiological signal, the first motion signal, and the second motion signal are collected by an integrated sensing device worn on the user; the integrated sensing device includes at least a heart rate sensor, an electromyography sensor, and a triaxial accelerometer.

[0010] In a second aspect, the present invention provides a motion state monitoring system based on high-precision sensing, used to implement the motion state monitoring method based on high-precision sensing described in the first aspect above; the system includes: an acquisition unit, used to acquire a user's physiological signal, a first motion signal, and a second motion signal; the physiological signal is used to characterize the user's muscle fatigue level and autonomic nervous system fatigue load; the first motion signal is used to characterize the user's motion state; the second motion signal is used to characterize the device's motion state; a first determination unit, used to determine the user's real-time physiological fatigue index based on the physiological signal; the real-time physiological fatigue index is used to characterize the user's current fatigue level based on physiological responses; a second determination unit, used to determine a wear-based reliability weight for evaluating the reliability of the physiological signal based on the first motion signal and the second motion signal; a third determination unit, used to determine a historical estimated fatigue index based on the second motion signal and historical motion data; the historical estimated fatigue index is used to characterize the user's fatigue level determined based on historical motion data; and a fusion unit, used to perform a weighted summation of the real-time physiological fatigue index and the historical estimated fatigue index based on the wear-based reliability weight to determine the user's fused fatigue index.

[0011] Thirdly, the present invention provides a computer-readable storage medium for a motion state monitoring device based on high-precision sensing. The device includes: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the motion state monitoring method based on high-precision sensing as described in the first aspect.

[0012] The present invention has the following beneficial effects: This invention simultaneously acquires physiological signals characterizing muscle fatigue and autonomic nervous system load, user-end motion signals, and device-end motion signals. It calculates a real-time physiological fatigue index and a wear-related reliability weight for each, and combines this with historical motion data to determine a historical estimated fatigue index. Finally, it weights and sums the real-time and historical indices to obtain a fused fatigue index. Compared to existing technologies that rely solely on current physiological signals for fatigue assessment, this approach fully considers the impact of sensor wear status on signal reliability. It utilizes the characteristic that device-end motion signals are unaffected by wear interference to correct user-end data, effectively avoiding assessment bias caused by poor sensor contact during exercise. This significantly improves the accuracy and robustness of exercise fatigue monitoring. It also solves the technical problem that changes in sensor-skin contact status can easily lead to a decrease in the quality of acquired physiological signals, resulting in low accuracy in exercise fatigue assessment in existing technologies. Attached Figure Description

[0013] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a flowchart illustrating a motion state monitoring method based on high-precision sensing provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the system architecture of a motion state monitoring system based on high-precision sensing provided in an embodiment of the present invention. Detailed Implementation

[0015] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the motion state monitoring method, device, and system based on high-precision sensing proposed by the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0016] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0017] The specific solutions of the motion state monitoring method, device and system based on high-precision sensing provided by the present invention are described in detail below with reference to the accompanying drawings.

[0018] Please see Figure 1 The diagram illustrates a motion state monitoring method based on high-precision sensing according to an embodiment of the present invention. The method includes the following steps S101-S105, which will be described in detail below.

[0019] S101. Acquire the user's physiological signals, first motion signals, and second motion signals.

[0020] Among them, physiological signals are used to characterize the user's muscle fatigue level and autonomic nervous system fatigue load; the first motion signal is used to characterize the user's motion state; and the second motion signal is used to characterize the device's motion state.

[0021] In one possible implementation, the user-worn integrated sensing device includes at least a heart rate sensor, an electromyography (EMG) sensor, and an accelerometer. This device collects physiological signals from the user at rest, including heart rate and EMG signals, and calculates corresponding baseline values ​​based on the collected resting signals for subsequent real-time signal normalization. Simultaneously, the accelerometer in the integrated sensing device collects a first motion signal from the user. During the user's actual exercise, the integrated sensing device continuously collects the user's physiological signals (including EMG and heart rate signals) and the first motion signal in real time. Furthermore, a second motion signal from the device is collected in real time using an accelerometer pre-configured on the fitness equipment. The collected signals undergo preprocessing, including rectification, smoothing filtering, noise reduction, and signal interval segmentation. The preprocessed data is divided and stored according to preset time windows, with each time window outputting a set of synchronized physiological signals, the first motion signal, and the second motion signal for subsequent steps.

[0022] For example, the integrated sensing device can be worn on the user's wrist, ankle, or waist, in close contact with the skin to ensure signal acquisition quality. Resting-state data acquisition is typically performed before exercise begins, while the user remains stationary for 1 to 3 minutes. During exercise, data is continuously output in fixed-duration windows (e.g., 5 seconds). Preprocessing is performed to remove motion artifacts and noise interference, ensuring signal availability.

[0023] S102. Determine the user's real-time physiological fatigue index based on physiological signals.

[0024] In one possible implementation, electromyography (EMG) and heart rate signals from physiological signals are used to calculate EMG-related factors to characterize muscle fatigue and heart rate-related factors to characterize autonomic nervous system fatigue, respectively. Specifically, an EMG fatigue factor is determined based on the change in the integral value of the EMG signal within each time window relative to the resting EMG integral baseline value. Simultaneously, a heart rate fatigue factor is determined based on the attenuation of the high-frequency and low-frequency components of the heart rate signal after frequency domain decomposition within each time window relative to their corresponding resting baseline values. Combining the EMG fatigue factor and the heart rate fatigue factor yields a real-time physiological fatigue index for that time window. This index characterizes the user's current level of fatigue based on physiological responses; a higher value indicates a higher level of fatigue.

[0025] S103. Based on the first motion signal and the second motion signal, determine the wear credibility weight for evaluating the credibility of the physiological signal.

[0026] In one possible implementation, the user's own motion intensity and the device's motion intensity are evaluated based on a first motion signal from the user and a second motion signal from the device. Both the first and second motion signals are acceleration signals. A pre-defined motion intensity analysis model (e.g., a recurrent neural network) converts the acceleration signals within each time window into corresponding motion intensity values. The difference between the user's motion intensity and the device's motion intensity is compared, and a wearing reliability weight is determined based on this difference. This weight is used to assess the contact state between the sensor and the user's skin surface within the current time window; a smaller difference indicates a better wearing condition, higher physiological signal reliability, and a larger wearing reliability weight value.

[0027] Understandably, the first motion signal is collected by an accelerometer worn by the user, directly reflecting the actual movement state of the user's limbs (such as the wrist, ankle, or waist), including changes in amplitude, frequency, and direction. The second motion signal is collected by an accelerometer pre-installed on the fitness equipment (such as a treadmill platform or a stationary bike frame), reflecting the vibration or motion state of the equipment itself. The key difference between the two is that the first motion signal is significantly affected by the user's wearing condition—if the sensor is loose or displaced, the signal may attenuate or become distorted; while the second motion signal is completely unaffected by the user's wearing condition, depending only on the equipment's own operating parameters (such as speed, incline, and resistance), and can serve as an objective benchmark for motion intensity. Therefore, using the difference between these two types of signals to evaluate the wear reliability weight has significant technical implications. Under ideal wearing conditions, the user's limb movement should be highly consistent with the equipment's movement (for example, the waveform trends of wrist acceleration and treadmill vibration acceleration during running are similar). At this point, the difference between the user's and the equipment's motion intensity is very small, indicating good sensor fit and high reliability of the physiological signal. Conversely, if the user wears the device improperly (e.g., the wristband is loose), the user's motion intensity may deviate significantly from the device's motion intensity, indicating a risk of physiological signal distortion. Based on this principle, by comparing the consistency of the two types of motion intensity, a quantitative assessment of the sensor wearing quality can be achieved, providing a weighting basis for subsequent fusion.

[0028] S104. Determine the historical estimated fatigue index based on the second motion signal and historical motion data.

[0029] In one possible implementation, a historical motion database is maintained to store data records of users wearing the device well during historical movements. Each record includes at least a second motion signal for a time window and its corresponding real-time physiological fatigue index. For the current time window, the interval of the wearing confidence weight determined in step S103 is first determined. When the weight is higher than a first threshold (for example, the first confidence threshold is 0.9, which is set empirically to indicate that when the wearing confidence weight reaches 90% or more, the sensor can be considered to be in good contact with the skin, and the physiological signal has high confidence), it indicates that the current wearing status is good. The second motion signal of the current window and its real-time physiological fatigue index are then stored as new historical data in the database to continuously enrich the historical samples. When the weight is lower than the first threshold but higher than the second threshold (for example, the second confidence threshold is 0.1, which is set based on experience, indicating that the signal is severely distorted when the confidence weight is lower than 10%, and the user should be prompted to wear the sensor again), it indicates that the current physiological signal is distorted to a certain extent. At this time, based on the second motion signal of the current window, the historical records most similar to the current second motion signal are retrieved from the historical motion database, and the real-time physiological fatigue index in the historical records is used as the historical estimated fatigue index of the current window.

[0030] For example, a sequence similarity analysis algorithm (such as dynamic time warping algorithm) is used to calculate the similarity between the current second motion signal and each historical second motion signal in the database, and the real-time physiological fatigue index corresponding to the historical record with the highest similarity is selected as the historical estimated fatigue index. For the initial stage of use or when there is not enough historical data, the real-time physiological fatigue index can be directly used as the historical estimated fatigue index, or only the real-time physiological signal can be relied upon for evaluation.

[0031] S105. Based on the wear-related confidence weight, the real-time physiological fatigue index and the historical estimated fatigue index are weighted and summed to determine the user's fusion fatigue index.

[0032] In one possible implementation, a wear-reliability weight is used as the fusion coefficient for the real-time physiological fatigue index, and the result of subtracting this weight from the unit value is used as the fusion coefficient for the historically estimated fatigue index. The real-time physiological fatigue index and the historically estimated fatigue index are then weighted and summed; the resulting weighted sum is the fused fatigue index for that time window. This fusion result integrates currently measurable physiological information with fatigue information estimated based on historical experience, and the weight allocation is matched to the current sensor-wearing status.

[0033] The technical solution provided by the above embodiments can bring at least the following beneficial effects: This embodiment simultaneously acquires physiological signals characterizing muscle fatigue and autonomic nervous system load, user-end motion signals, and device-end motion signals, calculates the real-time physiological fatigue index and wearing reliability weight respectively, and determines the historical estimated fatigue index by combining historical motion data. Finally, the real-time index and the historical index are weighted and summed to obtain the fused fatigue index. Compared with the existing technology that relies solely on the current physiological signals for fatigue assessment, this solution fully considers the impact of sensor wearing status on signal reliability, utilizes the characteristic that device-end motion signals are not affected by wearing interference to correct user-end data, effectively avoids assessment bias caused by poor sensor contact during exercise, and significantly improves the accuracy and robustness of exercise fatigue state monitoring. This solves the technical problem that the contact state between the sensor and the skin surface is prone to change, leading to a decrease in the quality of the collected physiological signals, thus resulting in low accuracy in the assessment of exercise fatigue state in the existing technology.

[0034] In one possible implementation, the physiological signals include electromyography (EMG) signals and heart rate signals. The process of determining the user's real-time physiological fatigue index based on the physiological signals can be specifically implemented through the following steps S201-S203, which will be described in detail below.

[0035] S201. Determine the electromyographic fatigue factor based on the integral value of the electromyographic signal within a preset time window relative to the baseline value of the electromyographic integral in the resting state.

[0036] In one possible implementation, for each time window, the integrated value of the preprocessed electromyographic (EMG) signal within that window is calculated; that is, the EMG amplitudes at all sampling points within the window are summed. This integrated value is then compared with a baseline EMG integral value collected from the user at rest before the start of exercise. The resulting ratio is the EMG fatigue factor for that window. This factor characterizes the degree of compensatory increase in electrical activity of the muscle relative to the resting state; a higher ratio indicates a higher degree of muscle fatigue.

[0037] Understandably, in practical applications, a user's heart rate and electromyography (EMG) signals exhibit normal physiological fluctuations at rest, with their integral values ​​both greater than zero. For healthy users, the frequency domain power integral value of the resting heart rate signal is typically in the range of 0.1 to 10 milliseconds squared, and the EMG integral value is typically in the range of 0.01 to 0.1 millivolts per second. Therefore, a denominator of zero would not occur under normal physiological conditions. If the resting baseline value is zero due to sensor malfunction, electrode detachment, or abnormal signal acquisition, it indicates that the acquisition device is in an abnormal working state. In this case, the system will stop calculating the fatigue index and issue a sensor malfunction or re-wearing prompt. The monitoring process will only resume after the user re-wears the device and successfully acquires a valid resting baseline value.

[0038] S202. Determine the heart rate fatigue factor based on the first attenuation factor and the second attenuation factor of the heart rate signal within a preset time window.

[0039] Wherein, the first attenuation factor is the ratio of the high-frequency component integral value of the heart rate signal within a preset time window to the high-frequency integral reference value in the resting state; the second attenuation factor is the ratio of the low-frequency component integral value of the heart rate signal within a preset time window to the low-frequency integral reference value in the resting state.

[0040] In one possible implementation, a frequency domain transformation is performed on the preprocessed heart rate signal within each time window to obtain the heart rate variability power spectrum. The power integral value within the high-frequency band is extracted and compared with the baseline power integral value of the same frequency band at rest. This ratio is then truncated using a minimum value function, i.e., compared to 1, and the smaller value is taken as the first attenuation factor. This factor characterizes the degree of preservation of parasympathetic regulatory activity relative to the resting level, and its value is limited to between 0 and 1. Similarly, the power integral value within the low-frequency band is extracted and compared with the baseline power integral value of the same frequency band at rest. This is also truncated using a minimum value function, and the smaller value between the ratio and 1 is taken as the second attenuation factor. This factor characterizes the degree of preservation of mixed sympathetic and parasympathetic regulatory activity relative to the resting level, and its value is also limited to between 0 and 1. The first attenuation factor and the second attenuation factor are multiplied to construct a heart rate fatigue factor. This factor is negatively correlated with the product value, that is, the more severe the attenuation (the smaller the product value), the larger the heart rate fatigue factor.

[0041] Understandably, in practical applications, a user's heart rate signal at rest exhibits normal physiological fluctuations, with its frequency domain power integral value always greater than zero. For healthy users, the high-frequency power integral value of the resting heart rate signal is typically in the range of 0.5 to 10 milliseconds squared, and the low-frequency power integral value is typically in the range of 0.5 to 20 milliseconds squared. Therefore, a denominator of zero would not occur under normal physiological conditions. If the resting baseline value is zero due to sensor malfunction, poor electrode contact, or abnormal signal acquisition, it indicates that the acquisition device is in an abnormal working state. In this case, the system will stop calculating the subsequent heart rate fatigue factor and issue a prompt to re-wear the device or check the sensor. The monitoring process will only resume after the user re-wears the device and successfully acquires a valid resting baseline value.

[0042] For example, the heart rate fatigue factor satisfies the following formula: in, is the heart rate fatigue factor for the k-th window, dimensionless, with a value range of [0.5, 1]. The larger the value, the higher the degree of fatigue based on heart rate. The first attenuation factor; The second attenuation factor; during fatigue and Both decrease, product The smaller the denominator, the smaller the fraction becomes. Enlarge It is positively correlated with the degree of fatigue. The constant 1 in the denominator has two important functions: ensuring that the denominator is always greater than or equal to 1, avoiding division by zero errors; and ensuring that the denominator is always greater than or equal to 1 under resting conditions. )hour Instead of 0 or 1, this aligns with physiological common sense, as heart rate variability has a baseline activity even at rest, thus avoiding extreme output values. The formula integrates two attenuation factors through a product; if either factor is very small (corresponding to severe inhibition of a certain autonomic nervous function), the product will also be small, making... A value close to 1 indicates high fatigue.

[0043] S203. Based on electromyographic fatigue factor and heart rate fatigue factor, determine the real-time physiological fatigue index.

[0044] In one possible implementation, the electromyographic fatigue factor and the heart rate fatigue factor are multiplied, and the product is the real-time physiological fatigue index for that time window. This index integrates information from both muscle fatigue and autonomic nervous system fatigue to comprehensively characterize the user's current level of fatigue based on physiological responses; a higher value indicates a higher overall level of fatigue.

[0045] Understandably, this step achieves synergistic enhancement of the two fatigue factors through multiplicative fusion rather than additive fusion. The real-time physiological fatigue index only increases significantly when both the muscular and cardiovascular levels exhibit fatigue characteristics; if only one indicator is abnormal while the other is normal, the product increase is limited, thus reducing the risk of false alarms from a single sensor.

[0046] The technical solution provided by the above embodiments can bring at least the following beneficial effects: This embodiment further defines the specific calculation method of the real-time physiological fatigue index, that is, the electromyographic fatigue factor is determined based on the integral value of the electromyographic signal, and the heart rate fatigue factor is determined based on the attenuation factors of the high-frequency and low-frequency components of the heart rate signal, and then the two are combined. The electromyographic signal reflects the decrease in muscle contraction efficiency and the increase in compensation, and the frequency domain analysis of the heart rate signal reflects the changes in autonomic nervous regulation function. Both reflect the degree of fatigue from the muscle level and the cardiovascular level, making the calculation of the real-time physiological fatigue index more comprehensive and scientific, and providing a reliable physiological basis for subsequent fusion assessment.

[0047] In one possible implementation, the process of determining the wear credibility weight for evaluating the credibility of physiological signals based on the first motion signal and the second motion signal can be specifically implemented through the following S301-S303, which will be described in detail below.

[0048] S301. Determine the motion intensity of the user terminal based on the first motion signal.

[0049] In one possible implementation, the first motion signal collected from the user within each time window is input into a preset motion intensity analysis model, which outputs a normalized motion intensity value. The first motion signal is collected by an accelerometer worn by the user, typically a triaxial acceleration signal; the resultant acceleration can be calculated first to eliminate directional influences. The motion intensity analysis model is used to establish the mapping relationship between the acceleration signal and motion intensity, and the output value quantifies the overall intensity of the user's motion within that time window.

[0050] For example, a motion intensity analysis model based on a Long Short-Term Memory (LSTM) network is used. This network is pre-trained with manually labeled acceleration signals of motion intensity. The input is triaxial acceleration time-series data within a time window, and the output is a motion intensity value between 0 and 1, where 0 represents stillness and 1 represents maximum motion intensity. The resultant acceleration is obtained by first calculating the modulus of the triaxial acceleration signal, and then input into the network for processing. During the motion, the motion intensity is output to the user in real time after each window ends, denoted as . .

[0051] S302. Determine the motion intensity at the device end based on the second motion signal.

[0052] In one possible implementation, for the second motion signal collected from the device within each time window, the same motion intensity analysis model as in step S301 is used. The second motion signal is input into the model, and the motion intensity value of the device is output. The second motion signal is collected by an accelerometer pre-configured on the fitness equipment (e.g., a treadmill or exercise bike) and is used to characterize the motion state of the equipment itself. Since the accelerometer is fixed to the equipment, the signal it collects is not affected by the user's wearing status and can objectively reflect the intensity of the equipment's operation.

[0053] For example, the vibration signal collected by the accelerometer on the treadmill is input into the same long short-term memory network model as in step S301, and the device-side motion intensity value is output, denoted as... For treadmills, the exercise intensity at the device end is highly correlated with parameters such as running speed and incline; for exercise bikes, it is related to parameters such as cycling resistance and cadence. This model shares the same network structure and parameters as the user-end exercise intensity model, ensuring that the exercise intensity output at both ends has consistent dimensions and value range.

[0054] S303. Determine the wear confidence weight based on the user's motion intensity and the device's motion intensity.

[0055] In one possible implementation, the degree of difference between the user's motion intensity and the device's motion intensity is calculated, and a wearing reliability weight is determined based on this difference. The wearing reliability weight is used to evaluate the contact state between the sensor and the user's skin surface within the current time window. Its value is negatively correlated with the difference between the user's motion intensity and the device's motion intensity; that is, the smaller the difference, the better the wearing condition, the higher the reliability of the physiological signal, and the larger the wearing reliability weight value.

[0056] For example, the wearable trust weight satisfies the following formula: in, is the wear credibility weight of the k-th window, which is dimensionless and takes a value range of [0, 1]. The larger the value, the better the sensor wear status and the more reliable the physiological signal. The motion intensity at the device end is calculated from the acceleration signal at the device end using a preset motion intensity analysis model (such as a long short-term memory artificial neural network). The user's motion intensity is calculated from the accelerometer signal worn by the user using the same model; This represents the absolute value of the difference between the two. The differences are converted into reliable weights; the smaller the difference, the closer the weight is to 1. Device-side motion is unaffected by the user's wearing condition, representing the true motion intensity; user-side motion intensity is affected by wearing tightness, displacement, etc. Consistency between the two indicates good sensor fit; differences... It is negatively correlated with wearing quality, so subtracting this difference from 1 yields a positively correlated confidence weight. The quality of the physiological signal in the current period was quantified, serving as the basis for the confidence level of subsequent weighted fusion.

[0057] The technical solution provided by the above embodiments can bring at least the following beneficial effects: This embodiment defines the method for determining the wearing reliability weight, that is, the motion intensity of the user end and the motion intensity of the device end are determined according to the first motion signal and the second motion signal respectively, and the wearing reliability weight is determined based on the negative correlation between the two. Since the motion signal of the device end is not affected by the user's wearing state, when the motion intensity of the user end and the device end are closer, it indicates that the sensor is in good contact with the skin and the physiological signal is highly reliable; conversely, it indicates that the wearing state is poor and the physiological signal is unreliable. This mechanism can objectively and quantitatively evaluate the quality of the physiological signal at each time period, providing an accurate reliability basis for subsequent dynamic adjustment of the fusion weight.

[0058] In one possible implementation, the process of determining the historical estimated fatigue index based on the second motion signal and historical motion data can be specifically implemented through the following S401-S403, which will be explained in detail below.

[0059] S401. When the wearer’s confidence weight is greater than or equal to the first confidence threshold, the real-time physiological fatigue index and the corresponding second motion signal are saved as a set of historical data for use in constructing or updating historical motion data.

[0060] In one possible implementation, during motion monitoring, for each time window, the relationship between the wearing confidence weight and a preset first confidence threshold is first determined. When the wearing confidence weight is greater than or equal to the first confidence threshold, it is considered that the sensor is in good contact with the user's skin surface within the current window, and the physiological signal has high confidence. At this time, the real-time physiological fatigue index of that window and the corresponding second motion signal (i.e., the device-side acceleration signal) are stored as a set of high-quality historical data in a historical motion database to enrich or update historical samples. Each record in this database contains the second motion signal of a time window and its associated real-time physiological fatigue index, for retrieval when the wearing condition is poor.

[0061] For example, the first confidence threshold is empirically set to 0.9, meaning that when the wearing confidence weight reaches 90% or higher, the current wearing condition is considered good. At this time, the real-time physiological fatigue index of the k-th window and the second motion signal of that window are jointly stored in the historical database. The database can be maintained using a queue structure, retaining only historical data from the most recent few movements (e.g., the most recent 10 movements or the most recent 1000 windows) to avoid unlimited data growth.

[0062] S402. When the wearing confidence weight is less than the first confidence threshold and greater than the second confidence threshold, based on the current second motion signal, retrieve historical second motion signals from historical motion data that meet the preset conditions for similarity with the current second motion signal.

[0063] Among them, the first confidence threshold is greater than the second confidence threshold.

[0064] In one possible implementation, when the wearing confidence weight is less than a first confidence threshold but greater than a second confidence threshold, it indicates that the physiological signal within the current window is distorted to a certain extent, but its confidence level is not low enough to necessitate forced re-wearing. In this case, the real-time physiological fatigue index is no longer directly accepted. Instead, based on the second motion signal (device-side acceleration signal) within the current window, the most similar historical second motion signal is retrieved from the historical motion database. The retrieval is based on the similarity of the signal waveform or motion pattern; historical records whose similarity meets preset conditions (e.g., highest similarity) are the retrieval results.

[0065] For example, the second confidence threshold is empirically set to 0.1. That is, when the confidence weight is below 10%, the signal is considered severely distorted, and the user should be prompted to wear the device again. When the weight is between 0.1 and 0.9, the historical retrieval mechanism is activated. A dynamic time warping algorithm is used to calculate the similarity between the current second motion signal and each historical second motion signal in the database, and the historical second motion signal with the highest similarity is selected as the retrieval result. The similarity threshold is preset to 0.7. If the highest similarity is still below this threshold, it is considered that there is no valid match, and in this case, it can revert to using only the real-time physiological fatigue index or issue a prompt.

[0066] S403. The historical real-time physiological fatigue index corresponding to the retrieved historical second motion signal is determined as the historical estimated fatigue index.

[0067] In one possible implementation, based on a historical second motion signal, a historical real-time physiological fatigue index associated with and stored in a historical motion database is extracted. This historical real-time physiological fatigue index is then used as the historical estimated fatigue index for the current window, and subsequently weighted and fused with the real-time physiological fatigue index. This index characterizes the degree of fatigue historically exhibited by the user under similar current motion patterns.

[0068] For example, suppose the window number corresponding to the retrieved historical second motion signal is l, and the historical real-time physiological fatigue index of this window is . Then the historical estimated fatigue index of the current window is equal to If the retrieval result of step S402 is multiple historical signals (e.g., selecting the top three with the highest similarity), the average of their corresponding historical real-time physiological fatigue indices can be used as the historical estimated fatigue index. For initial runs or cases with insufficient historical data, the historical estimated fatigue index can be set to 0.5 by default, or the real-time physiological fatigue index can be used directly.

[0069] The technical solution provided by the above embodiments can bring at least the following beneficial effects: This embodiment defines the determination logic of the historical estimated fatigue index, and performs graded processing of the wearing status by setting a first confidence threshold and a second confidence threshold. When the wearing confidence weight is high, the current data is stored in the historical motion database as a high-quality historical sample to realize the continuous accumulation and updating of historical data; when the wearing confidence weight is in the middle range, the historical fatigue index corresponding to similar motion patterns is retrieved from the historical database based on the current device motion signal as the estimate. This graded strategy not only ensures the quality of historical data, but also can use historical experience to compensate when physiological signals are unreliable, effectively improving the system's evaluation capability under poor wearing conditions.

[0070] In one possible implementation, the process of retrieving historical second motion signals that meet the preset conditions of similarity with the current second motion signal from historical motion data when the wearing confidence weight is less than the first confidence threshold and greater than the second confidence threshold can be specifically implemented through the following S501-S502, which will be explained in detail below.

[0071] S501. Based on the sequence similarity analysis algorithm, determine the similarity between the current second motion signal and each historical second motion signal in the historical motion data.

[0072] In one possible implementation, for the current second motion signal, a sequence similarity analysis algorithm is used to calculate the similarity between this signal and each historical second motion signal stored in the historical motion database. The sequence similarity analysis algorithm measures the degree of matching between two time-series signals in waveform morphology, outputting a quantization index; the larger the index value, the more similar the two signals are. During the calculation, the second motion signal needs to undergo modulus calculation of the combined acceleration to transform the triaxial acceleration into a one-dimensional time-series sequence, thus eliminating the influence of direction.

[0073] For example, the dynamic time warping algorithm is used as the sequence similarity analysis algorithm. Let the current second motion signal sequence be... The length is T; the second motion signal sequence of the l-th record in the historical database is The sequence has a length of T'. The Dynamic Time Warping algorithm finds the optimal alignment path between two sequences and calculates the cumulative distance. The smaller the distance value, the more similar the sequences are. For easier comparison, the distance value can be converted into a similarity score using a reciprocal transformation. The converted similarity score ranges from 0 to 1, where 1 indicates identical sequences and 0 indicates completely different sequences. This process is repeated sequentially. Similarity to all historical signal sequences in the database.

[0074] For example, distance values ​​can be converted into similarity using the following formula: in, This is the second motion signal sequence (combined acceleration timing) of the device at the current time window. For the first time in the historical database The second motion signal sequence recorded on the device side. The cumulative distance calculated by the dynamic time warping algorithm; The distance value is converted into a similarity score by using the inverse transformation. This transformation maps the distance to (0,1], when DTW=0. (Completely the same); when DTW approaches infinity, Approaching 0 (completely different).

[0075] S502. Determine the second motion signal with the highest similarity as the historical second motion signal.

[0076] In one possible implementation, all historical records are sorted from highest to lowest similarity based on the similarity between the current second motion signal and each historical second motion signal. The second motion signal from the historical record with the highest similarity is selected as the retrieved historical second motion signal for subsequent extraction of the corresponding historical real-time physiological fatigue index. If multiple historical records have the same highest similarity, any one can be selected, or the average fatigue index of the multiple historical second motion signals with the highest similarity can be taken.

[0077] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment defines a specific method for retrieving historical data when the wearer's trusted weight is in the middle range. Specifically, it uses a sequence similarity analysis algorithm to calculate the similarity between the current device's motion signal and signals from different historical time periods, and selects the second historical motion signal with the highest similarity as the retrieval result. Through sequence matching techniques such as dynamic time warping, the historical motion segment most similar to the current motion pattern can be accurately identified, thus ensuring a high correlation between the retrieved historical estimated fatigue index and the user's current actual fatigue state, further guaranteeing the accuracy of the fusion results.

[0078] In one possible implementation, the process of determining the user's fused fatigue index by weighted summation of the real-time physiological fatigue index and the historical estimated fatigue index based on the wearer's trusted weight can be specifically implemented through the following S601-S602, which will be explained in detail below.

[0079] S601. Based on the wear confidence weight, determine the first weight of the real-time physiological fatigue index and the second weight of the historical estimated fatigue index.

[0080] In one possible implementation, based on the wear-reliability weight, a first weight corresponding to the real-time physiological fatigue index and a second weight corresponding to the historical estimated fatigue index are determined. The first weight is positively correlated with the wear-reliability weight; that is, the better the wearing condition and the higher the reliability of the physiological signal, the greater the proportion of the real-time physiological fatigue index in the fusion process. The second weight is negatively correlated with the wear-reliability weight; that is, the worse the wearing condition and the lower the reliability of the physiological signal, the greater the proportion of the historical estimated fatigue index in the fusion process. Both weights range from 0 to 1, and the sum of the two weights equals 1, to ensure a smooth transition between the two extreme cases in the fusion result. In another possible implementation, the wear-reliability weight is directly used as the first weight of the real-time physiological fatigue index, i.e., the first weight equals the wear-reliability weight; simultaneously, the unit value of 1 minus the wear-reliability weight is used as the second weight of the historical estimated fatigue index.

[0081] For example, when the wearing confidence weight is 0.8, the first weight is 0.8 and the second weight is 0.2; when the wearing confidence weight is 0.3, the first weight is 0.3 and the second weight is 0.7. When the wearing confidence weight is 1 (wearing perfectly), the first weight is 1 and the second weight is 0, and the fusion result is exactly equal to the real-time physiological fatigue index; when the wearing confidence weight is 0 (wearing completely ineffective), the first weight is 0 and the second weight is 1, and the fusion result is exactly equal to the historical predicted fatigue index. The sum of the two weights is always 1, ensuring that the fusion result is always a convex combination of the real-time physiological fatigue index and the historical predicted fatigue index, with stable values ​​and a smooth transition between the two extreme cases.

[0082] S602. Determine the fusion fatigue index based on the real-time physiological fatigue index and its first weight, the historical estimated fatigue index and its second weight.

[0083] In one possible implementation, based on a first weight and a second weight, the real-time physiological fatigue index is multiplied by the first weight to obtain a first product, and the historically estimated fatigue index is multiplied by the second weight to obtain a second product. The first and second products are then added together to obtain the fused fatigue index for the current time window. This fused index integrates currently measurable physiological information and fatigue information estimated based on historical experience, serving as the final output assessment result of exercise fatigue status.

[0084] For example, the fusion fatigue index satisfies the following formula: in, Let f be the fusion fatigue index of the k-th window, which is dimensionless, and let f be the user's motion fatigue state in the final output. The wearable confidence weight has a value range of [0, 1]. Real-time physiological fatigue index; For historical fatigue index prediction, the value range and Quite good; when worn in good condition ( When the result is close to 1), the fusion result mainly depends on real-time physiological data, because the physiological signal has high reliability at this time.

[0085] When the wearing condition is poor ( When the weights are close to 0, the fusion result relies almost entirely on historical prediction data, avoiding interference from unreliable physiological signals; the sum of the two weights is 1, ensuring a smooth transition between the two extreme cases; this weighting strategy is a form of linear interpolation with a clear physical meaning: data with higher credibility is given higher weights. It integrates currently measurable physiological information with historical experience and knowledge, and can still output reliable fatigue assessment results even when signal quality deteriorates.

[0086] The technical solution provided by the above embodiments can bring at least the following beneficial effects: This embodiment defines a specific method for obtaining the fused fatigue index through weighted summation, namely, determining the first weight of the real-time physiological fatigue index and the second weight of the historically estimated fatigue index based on the wearing reliability weight, wherein the first weight is positively correlated with the wearing reliability weight, and the second weight is negatively correlated with the wearing reliability weight. This weight allocation mechanism ensures that when the sensor is worn well and the physiological signal reliability is high, the fusion result relies more on real-time physiological data; when the wearing condition is poor and the physiological signal reliability is low, the fusion result refers more to historically estimated data. This adaptive weighted fusion strategy fully leverages the complementary advantages of the two indices, enabling the final output fused fatigue index to maintain high accuracy under different wearing conditions.

[0087] In one possible implementation, the operating state of the equipment can be adjusted according to the equipment operating parameters corresponding to the target fatigue index range. This process can be specifically implemented through the following S701-S703, which will be described in detail below.

[0088] S701. Obtain multiple preset fatigue index ranges.

[0089] One possible implementation involves pre-setting multiple fatigue index intervals, each corresponding to a different fatigue level, and associating each interval with a set of equipment operating parameters. The fatigue index intervals are divided based on the range of values ​​for the integrated fatigue index, which can be set according to exercise physiology experience or experimental statistical data. These intervals are used to classify the current user's fatigue state so that differentiated equipment control strategies can be adopted subsequently.

[0090] For example, the real-time physiological fatigue index is the product of the electromyographic fatigue factor and the heart rate fatigue factor. The electromyographic fatigue factor is the ratio of the current window's electromyographic integral value to the resting baseline value. At rest, the electromyographic fatigue factor is approximately 1. During mild exercise, muscle compensation is slight, and the electromyographic fatigue factor is usually between 1 and 2. When moderate to high-intensity exercise leads to muscle fatigue, the integral value can rise to 2 to 4 times the resting value. During extreme exercise or high-intensity strength training, the electromyographic fatigue factor can reach 5 to 8, and in very rare extreme cases, it may exceed 10. In actual exercise, the common range of the electromyographic fatigue factor is 1 to 8, rarely exceeding 10. The heart rate fatigue factor ranges from [0.5, 1]. Considering that the electromyographic integral value is unlikely to exceed 20 times the resting value in actual exercise, and the maximum heart rate fatigue factor is 1, the theoretical upper limit of the real-time physiological fatigue index is approximately 20. In typical exercise scenarios, the measured values ​​mostly fall between 0.5 and 20. Considering rare anomalies (such as the sensor's intense movement causing the electromyographic fatigue factor to briefly exceed 20), the upper limit can be extended to 30 to cover all possibilities. Therefore, the range of the fused fatigue index can be divided into four intervals: the first interval is 0 to 10, corresponding to a normal exercise mode, indicating low user fatigue and the ability to maintain normal exercise intensity; the second interval is 10 to 20, corresponding to a mild fatigue mode, indicating the user is beginning to show signs of fatigue; the third interval is 20 to 30, corresponding to a moderate fatigue mode, indicating a high level of user fatigue; and the fourth interval is above 30, corresponding to a severe fatigue mode, indicating the user is highly fatigued and needs rest. Each interval is pre-associated with a set of equipment operating parameters, such as the treadmill's speed and incline, or the resistance level of a stationary bike. The interval boundary values ​​can be personalized based on individual user differences (such as age and fitness level).

[0091] S702. Determine the target fatigue index range to which the fusion fatigue index belongs.

[0092] In one possible implementation, based on the fused fatigue index, the index is compared with multiple preset fatigue index intervals to determine which interval the index falls into, and that interval is then designated as the target fatigue index interval. If the fused fatigue index happens to be located on the interval boundary, it can be processed according to preset rules (e.g., belonging to a lower or higher interval).

[0093] S703. Adjust the operating status of the equipment according to the equipment operating parameters corresponding to the target fatigue index range.

[0094] In one possible implementation, based on a target fatigue index range, a set of equipment operating parameters associated with that range is read from a preset configuration, and the current operating state of the equipment is adjusted to the state specified by those parameters. The adjustment operation may include changing the equipment's movement speed, resistance level, slope, vibration intensity, etc., and may also include outputting prompts (such as voice reminders or screen displays). The magnitude and rate of adjustment can be set according to safety regulations to avoid sudden changes that could cause user discomfort.

[0095] For example, for treadmills, when the target zone is the first zone (normal exercise mode, fatigue index 0-10), maintain the default speed (e.g., 8 km / h) determined based on the user's body type and initial settings. When the target zone is the second zone (mild fatigue mode, 10-20), reduce the speed by 10% to 7.2 km / h and display a message on the screen saying "You are slightly fatigued, it is recommended to slow down appropriately." When the target zone is the third zone (moderate fatigue mode, 20-30), further reduce the speed by 20% to 6.4 km / h and issue a voice reminder saying "You are moderately fatigued, please adjust your breathing." When the target zone is the fourth zone (severe fatigue mode, 30 and above), gradually reduce the speed to 0 and prompt "You are severely fatigued, please stop exercising and rest." For exercise bikes, adjust the resistance level accordingly.

[0096] The technical solution provided in the above embodiments can bring at least the following beneficial effects: This embodiment applies the fusion fatigue index to equipment operation control. By preset multiple fatigue index ranges and their corresponding equipment operating parameters, the operating state of the equipment is automatically adjusted according to the range to which the fusion fatigue index belongs. This enables the exercise equipment to dynamically adjust the exercise intensity or resistance according to the user's real-time fatigue level. When the user is slightly fatigued, the intensity is appropriately reduced; when the user is moderately fatigued, the speed is further reduced; and when the user is severely fatigued, a rest reminder is given. This achieves intelligent exercise intervention, effectively preventing users from developing health risks due to excessive exercise, and improving the safety of the exercise equipment and the user experience.

[0097] In one possible implementation, physiological signals, a first motion signal, and a second motion signal are collected by an integrated sensing device worn on the user; the integrated sensing device includes at least a heart rate sensor, an electromyography sensor, and a triaxial accelerometer.

[0098] The technical solution provided by the above embodiments can bring at least the following beneficial effects: This embodiment limits the physiological signals, the first motion signal, and the second motion signal to be collected by an integrated sensing device, which includes at least a heart rate sensor, an electromyography (EMG) sensor, and a triaxial accelerometer. By integrating multiple sensors into the same wearable device (such as a ring-shaped wearing band), the user's heart rate, EMG, and acceleration signals can be collected synchronously, while the acceleration signal from the device itself can be obtained through the device's own sensors. This integrated acquisition method simplifies hardware deployment, ensures the temporal synchronization of multi-source signals, facilitates subsequent signal fusion processing, and reduces the complexity of the system and the barrier to entry for use.

[0099] Please see Figure 2 This document illustrates a schematic diagram of the system architecture of a motion state monitoring system 200 based on high-precision sensing, provided in an embodiment of the present invention. This system implements the aforementioned motion state monitoring method based on high-precision sensing. The system includes: an acquisition unit 201, used to acquire a user's physiological signals, a first motion signal, and a second motion signal; the physiological signals characterize the user's muscle fatigue level and autonomic nervous system fatigue load; the first motion signal characterizes the user's motion state; the second motion signal characterizes the device's motion state; a first determination unit 202, used to determine the user's real-time physiological fatigue index based on the physiological signals; the real-time physiological fatigue index characterizes the user's current fatigue level based on physiological responses; a second determination unit 203, used to determine a wear-based reliability weight for evaluating the reliability of the physiological signals based on the first and second motion signals; a third determination unit 204, used to determine a historical estimated fatigue index based on the second motion signal and historical motion data; the historical estimated fatigue index characterizes the user's fatigue level determined based on historical motion data; and a fusion unit 205, used to perform a weighted summation of the real-time physiological fatigue index and the historical estimated fatigue index based on the wear-based reliability weight to determine the user's fused fatigue index.

[0100] The technical solution provided by the above embodiments can bring at least the following beneficial effects: This embodiment provides a motion state monitoring system based on high-precision sensing that is completely corresponding to the above method. Through the modular design of the acquisition unit, the first determination unit, the second determination unit, the third determination unit, and the fusion unit, the system performs steps such as signal acquisition, real-time physiological fatigue index determination, wearable confidence weight determination, historical estimated fatigue index determination, and weighted fusion, respectively. This system transforms the method into a functional module architecture, which can efficiently and stably realize motion state monitoring, is easy to embed into various smart fitness devices or wearable devices, and has good practicality and scalability.

[0101] In one possible implementation, the present invention provides a motion state monitoring device based on high-precision sensing, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the above-described motion state monitoring method based on high-precision sensing.

[0102] The technical solutions provided in the above embodiments can bring at least the following beneficial effects: This embodiment provides a motion state monitoring device based on high-precision sensing, including a memory and a processor. The processor executes the stored computer program to realize the above-mentioned motion state monitoring method based on high-precision sensing. As a physical product, this device can be directly integrated into sports equipment or wearable devices such as treadmills, exercise bikes, and smart bracelets, realizing the technical solution of the present invention in hardware form.

[0103] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0104] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A motion state monitoring method based on high-precision sensing, characterized in that, The method includes: Acquire the user's physiological signals, a first motion signal, and a second motion signal; the physiological signals are used to characterize the user's muscle fatigue level and autonomic nervous system fatigue load; the first motion signal is used to characterize the user's motion state; the second motion signal is used to characterize the device's motion state. Based on the physiological signals, the user's real-time physiological fatigue index is determined; the real-time physiological fatigue index is used to characterize the user's current level of fatigue based on physiological responses. Based on the first motion signal and the second motion signal, a wear confidence weight is determined for evaluating the confidence of the physiological signal; Based on the second motion signal and historical motion data, a historical estimated fatigue index is determined; the historical estimated fatigue index is used to characterize the user's fatigue level determined based on historical motion data. Based on the wear confidence weight, the real-time physiological fatigue index and the historical estimated fatigue index are weighted and summed to determine the user's fusion fatigue index.

2. The motion state monitoring method based on high-precision sensing according to claim 1, characterized in that, The physiological signals include: electromyography signals and heart rate signals; determining the user's real-time physiological fatigue index based on the physiological signals includes: Based on the integral value of the electromyographic signal within a preset time window, relative to the baseline value of the electromyographic integral in the resting state, the electromyographic fatigue factor is determined. Based on the first attenuation factor and the second attenuation factor of the heart rate signal within the preset time window, a heart rate fatigue factor is determined; the first attenuation factor is the ratio of the high-frequency component integral value of the heart rate signal within the preset time window to the high-frequency integral reference value in the resting state; the second attenuation factor is the ratio of the low-frequency component integral value of the heart rate signal within the preset time window to the low-frequency integral reference value in the resting state. The real-time physiological fatigue index is determined based on the electromyographic fatigue factor and the heart rate fatigue factor.

3. The motion state monitoring method based on high-precision sensing according to claim 1, characterized in that, The step of determining the wear reliability weight for evaluating the reliability of the physiological signal based on the first motion signal and the second motion signal includes: The user's motion intensity is determined based on the first motion signal; The motion intensity at the device end is determined based on the second motion signal; The wear confidence weight is determined based on the user's motion intensity and the device's motion intensity.

4. The motion state monitoring method based on high-precision sensing according to claim 1, characterized in that, The step of determining the historical estimated fatigue index based on the second motion signal and historical motion data includes: When the wear confidence weight is greater than or equal to the first confidence threshold, the real-time physiological fatigue index and the corresponding second motion signal are saved as a set of historical data for the purpose of constructing or updating the historical motion data. When the wearable confidence weight is less than the first confidence threshold and greater than the second confidence threshold, based on the current second motion signal, a historical second motion signal whose similarity to the current second motion signal meets a preset condition is retrieved from the historical motion data; the first confidence threshold is greater than the second confidence threshold. The historical real-time physiological fatigue index corresponding to the retrieved historical second motion signal is determined as the historical estimated fatigue index.

5. The motion state monitoring method based on high-precision sensing according to claim 4, characterized in that, The step of retrieving historical second motion signals from the historical motion data that satisfy a preset condition in similarity to the current second motion signal based on the current second motion signal includes: Based on the sequence similarity analysis algorithm, the similarity between the current second motion signal and each historical second motion signal in the historical motion data is determined respectively; The second motion signal with the highest similarity is determined as the historical second motion signal.

6. The motion state monitoring method based on high-precision sensing according to claim 1, characterized in that, The step of determining the user's fused fatigue index by weighting and summing the real-time physiological fatigue index and the historical estimated fatigue index according to the wear confidence weight includes: Based on the wear-reliability weight, a first weight for the real-time physiological fatigue index and a second weight for the historical estimated fatigue index are determined; the first weight is positively correlated with the wear-reliability weight; the second weight is negatively correlated with the wear-reliability weight. The fusion fatigue index is determined based on the real-time physiological fatigue index and its first weight, the historical estimated fatigue index and its second weight.

7. The motion state monitoring method based on high-precision sensing according to claim 1, characterized in that, The method further includes: Obtain multiple preset fatigue index intervals; each fatigue index interval corresponds to a set of equipment operating parameters. Determine the target fatigue index range to which the fusion fatigue index belongs; The operating status of the equipment is adjusted according to the equipment operating parameters corresponding to the target fatigue index range.

8. The motion state monitoring method based on high-precision sensing according to claim 1, characterized in that, The physiological signals, the first motion signal, and the second motion signal are collected by an integrated sensing device worn by the user. The integrated sensing device includes at least: a heart rate sensor, an electromyography sensor, and a triaxial accelerometer.

9. A motion state monitoring system based on high-precision sensing, characterized in that, The system includes: The acquisition unit is used to acquire the user's physiological signals, a first motion signal, and a second motion signal; the physiological signals are used to characterize the user's muscle fatigue level and autonomic nervous system fatigue load; the first motion signal is used to characterize the user's motion state; and the second motion signal is used to characterize the device's motion state. The first determining unit is configured to determine the user's real-time physiological fatigue index based on the physiological signal; the real-time physiological fatigue index is used to characterize the user's current level of fatigue based on physiological response. The second determining unit is used to determine a wear confidence weight for evaluating the confidence of the physiological signal based on the first motion signal and the second motion signal. The third determining unit is used to determine the historical estimated fatigue index based on the second motion signal and historical motion data; the historical estimated fatigue index is used to characterize the fatigue level of the user determined based on the historical motion data. The fusion unit is used to perform a weighted summation of the real-time physiological fatigue index and the historical estimated fatigue index based on the wear confidence weight, so as to determine the user's fused fatigue index.

10. A motion state monitoring device based on high-precision sensing, characterized in that, include: Memory, used to store computer programs; A processor, when executing the computer program, implements the steps of the motion state monitoring method based on high-precision sensing as described in any one of claims 1 to 8.

Images