CN-122004782-A - Multi-muscle group fatigue detection algorithm based on myocardial bimodal signal time sequence phase difference accumulation

CN122004782ACN 122004782 ACN122004782 ACN 122004782ACN-122004782-A

Abstract

The invention relates to the technical field of physiological state monitoring, in particular to a multi-muscle group fatigue detection algorithm based on sequential phase difference accumulation of myocardial bimodal signals, which comprises the following steps of acquiring surface electromyographic signal sequences and electrocardiosignal sequences of a plurality of target muscles which are synchronously acquired, calculating multi-muscle group joint energy values and multi-muscle group cooperative non-scheduling indexes based on root mean square value sequences, identifying and extracting heart acceleration micro-events based on RR intervals, constructing individual reference values based on sequential phase differences, integrating and accumulating the phase differences exceeding the reference values to form a fatigue accumulation index, and judging the fatigue state trigger through the index. The algorithm accurately exposes implicit compensatory fatigue, greatly improves the real-time response capability of the algorithm, can effectively eliminate non-acting heart rate interference, avoids individual calibration, has the robustness of engineering landing, greatly reduces the calibration cost before industrial application, and remarkably improves the universality of the algorithm and the engineering deployment convenience.

Inventors

CAI YONGQING
ZHANG CHI
XU WEILIN
LI HAIYANG
TANG HAO
TAN QI
CHEN SHAOFENG

Assignees

合肥工业大学

Dates

Publication Date: 20260512
Application Date: 20260330

Claims (10)

1. The multi-muscle group fatigue detection algorithm based on the accumulation of the time sequence phase difference of the myocardial bimodal signals is characterized by comprising the following steps: step S1, acquiring surface electromyographic signal sequences and electrocardiosignal sequences of a plurality of target muscles which are synchronously acquired; S2, carrying out sliding window processing on the surface electromyographic signal sequences, extracting root mean square value sequences of all muscles, and calculating multi-muscle group joint energy values and multi-muscle group coordination failure indexes based on the root mean square value sequences; step S3, identifying and extracting muscle force micro-events based on the combined energy value and the cooperative non-scheduling index; Step S4, identifying and extracting cardiac acceleration micro-events based on RR intervals in the electrocardiosignal sequence; step S5, performing time axis matching on the muscle force-generating micro-event and the heart acceleration micro-event, and calculating time sequence phase difference between the matched event pairs; S6, constructing an individual reference value based on the time sequence phase difference, and integrating and accumulating the phase difference exceeding the reference value to form a fatigue accumulation index; and when the fatigue accumulation index exceeds a preset threshold value, judging that the fatigue state is triggered.
2. The multi-muscle group fatigue detection algorithm based on the accumulation of the time-series phase differences of the myocardial bimodal signals according to claim 1, wherein the formula for calculating the joint energy value in the step S2 is: ; Wherein, the Is a joint root mean square value; the root mean square value of the ith muscle is obtained, N is the total number of the muscles/the number of channels; natural logarithm operators; The smoothing factor/positive definite small constant is 10 −9 , and exp is an exponential reduction operator.
3. The multi-muscle group fatigue detection algorithm based on the accumulation of the myocardial bimodal signal time series phase difference according to claim 1, wherein the method for calculating the cooperative unscheduled operation in the step S2 is as follows: Firstly, calculating a first-order difference of each path of electromyographic signal sequence to obtain a force change rate sequence of each muscle; Then, at the same time section, calculating standard deviation of all muscle force change rates, and defining the standard deviation sequence as a coordinated failure schedule, wherein a specific calculation formula is as follows: ; wherein i is a muscle channel index, RMSi is a root mean square value of the ith muscle in the current time window, dRMSi/dt represents that a first-order difference is obtained on the RMS sequence of the ith muscle, sigma is standard deviation operation, and MSD (t) is a finally output multi-muscle group cooperative non-scheduling index.
4. The multi-muscle group fatigue detection algorithm based on the accumulation of the time-series phase differences of the myocardial bimodal signals according to claim 1, wherein the method for detecting the muscle force micro-event in the step S3 is as follows: Calculating a dynamic baseline of the combined energy value and a dynamic baseline of the collaborative non-scheduling in real time; when the measured combined energy value is greater than 1.5 times of the dynamic baseline of the combined energy value and the cooperative non-scheduled dynamic baseline is greater than 0.5 times of the cooperative non-scheduled dynamic baseline, judging that the current moment is a muscle force micro-event, wherein the specific judgment formula is as follows: ; The EEMG (t) is a binary indication function, RMSjoint (t) is a multi-muscle group joint root mean square energy value at the current moment, RMSbase is a preset healthy segment joint energy mean value, MSD (t) is a coordinated out-of-schedule at the current moment, MSDbase is a preset healthy segment coordinated out-of-schedule mean value, and n is a logic intersection operation.
5. The multi-muscle group fatigue detection algorithm based on the accumulation of the time-series phase differences of the myocardial bimodal signals according to claim 1, wherein the method for detecting the cardiac acceleration micro-event in the step S4 is as follows: Calculating an instantaneous heart rate and a change rate thereof based on the RR interval; When the rising amplitude of the instantaneous heart rate between adjacent sampling points exceeds a preset heart rate sudden rise threshold, judging the current moment as a heart acceleration micro-event, wherein the specific calculation formula is as follows: ; Wherein, the (T ') is a heart acceleration micro-event indication function, HR (t') is an instantaneous heart rate value calculated based on RR intervals, dHR (t ')/dt' is a first order difference of heart rate, and 5 is a preset physiological threshold value, which represents that the heart rate is accelerated more than 5 times per minute in 1 second, and the unit is bpm/s.
6. The multi-muscle group fatigue detection algorithm based on the accumulation of the myocardial bimodal signal time-series phase difference according to claim 1, wherein the calculating of the central muscle event time-series phase difference in the step S5 comprises the specific steps of matching each muscle force-generating micro-event with the first heart acceleration micro-event in a preset time window after the occurrence on a time axis, and the specific calculation formula is as follows: ; Wherein, the A set of myocardial events that are valid pairs of the kth pair; A time stamp representing the occurrence of a kth significant muscle development micro-event for the detected moment of onset of muscle development; 3s is a time sequence association window span, and the longest effective delay of the corresponding relation in physiology is defined; after matching, forming effective event pairs, and calculating the time difference between the occurrence time of the heart acceleration micro-event and the occurrence time of the muscle stress micro-event in each effective event pair to be used as a myocardial event time sequence phase difference, wherein the specific calculation formula is as follows: ; Wherein K is the sequence number of successfully matched myocardial event pairs, and phi K is the phase delay difference, namely the physiological delay time of the heart for performing vascular pumping compensation response to the muscle force action.
7. The multi-muscle group fatigue detection algorithm based on the accumulation of the myocardial bimodal signal time series phase difference according to claim 1, wherein the step S6 further comprises the steps of constructing and determining a fatigue accumulation index, specifically: establishing individual health state reference phase delay based on a plurality of myocardial event time sequence phase differences at the initial stage of operation; For each myocardial event time sequence phase difference generated subsequently, calculating an excess part exceeding the reference phase delay, and integrating and accumulating all the excess parts to obtain a fatigue accumulation index; judging whether the fatigue accumulation index exceeds a preset global threshold, if so, judging that the fatigue state is triggered.
8. The multi-muscle group fatigue detection algorithm based on the accumulation of myocardial bimodal signal time series phase differences according to claim 7, wherein the method for establishing the individual health state reference phase delay in step S6 is as follows: and taking the time sequence phase differences of a plurality of myocardial events corresponding to the effective event pairs of the front part events, and calculating the average value of the time sequence phase differences as the reference phase delay.
9. The multi-muscle group fatigue detection algorithm based on the accumulation of myocardial bimodal signal time series phase difference according to claim 8, wherein the formula for calculating the fatigue accumulation index FCI in step S6 is: ; Wherein, the A current value for the fatigue accumulation index; I.e. the current delay; bias for individual health benchmarks; is a nonlinear rectification operator/lower limit cutoff function.
10. The multi-muscle group fatigue detection algorithm based on the accumulation of the time-series phase differences of the myocardial bimodal signals according to claim 8, wherein the global threshold in the step S6 is a fixed constant for directly comparing and outputting the fatigue trigger signals.

Description

Multi-muscle group fatigue detection algorithm based on myocardial bimodal signal time sequence phase difference accumulation Technical Field The invention relates to the technical field of physiological state monitoring, in particular to a multi-muscle group fatigue detection algorithm based on myocardial bimodal signal time sequence phase difference accumulation. Background In the field of human body motion state monitoring and fatigue detection, surface electromyographic signals (sEMG) and heart rate (HR/HRV) are two physiological signals which are most widely applied at present, and important data support is provided for fatigue evaluation from two dimensions of local muscle electrical activity and overall cardiovascular load. Analysis methods based on surface electromyographic signals are conventional means for detecting muscle fatigue. The prior art usually quantifies muscle strength by extracting Root Mean Square (RMS) values or reflects the fatigue characteristics of the left shift of the muscle spectrum by extracting median frequency (MDF). The method has the advantages of high sampling frequency and high response speed, and can reflect the instant acting state of local muscles in real time. However, in a practical complex work scenario, human action often involves the cooperative work of multiple muscles. When one active muscle is fatigued, the nervous system instinctively mobilizes other muscles around to compensate for the force to maintain the overall output. In this case, if only a single muscle is monitored, or the energy of a plurality of muscles is simply subjected to arithmetic average, the essential characteristic that the local muscles are fatigued is covered up because the total energy of the system is not changed greatly, so that the algorithm cannot capture the key information that the cooperative work mode of the muscles is disturbed sharply. In addition, the traditional myoelectricity characteristic index is sensitive to absolute amplitude, and in actual deployment, calibration test of maximum autonomous contractility (MVC) is needed to be carried out in advance for each individual, so that calibration cost of engineering application is increased. Fatigue status is assessed from another angle based on an analytical method of heart rate variability. The heart rate variation can well reflect the overall functional load of operators and the macro-regulation state of an autonomic nervous system, and the signal has strong myoelectric noise interference resistance. However, there are inherent limitations to using it as a real-time fatigue detection indicator: On the one hand, the blood flow pumping response of the cardiovascular system itself has physiological delays, and the heart rate rise after muscle work does not occur instantaneously, on the other hand, conventional heart rate variability indices (such as LF/HF ratio in frequency domain analysis) must be mathematically dependent on a long time window (typically 1 to 5 minutes) to obtain stable statistics. Although this way of calculating the long window accumulation can smooth short term fluctuations, it necessarily results in significant detection delays in short term, fast-running or immediate feedback-required scenarios. More importantly, the reason why the heart rate rise cannot be effectively distinguished simply by heart rate fluctuation, namely whether the heart rate rise is caused by real physical work requirements of muscles or by non-metabolic factors such as emotion fluctuation, body position change or environmental stimulus, is simply caused, so that the fatigue detection algorithm based on the heart rate is insufficient in the practical application. In summary, how to use the high real-time feedback advantage of the electromyographic signals and exert the physiological value of the cardiovascular signals to reflect the overall central load, overcome the dual technical defects of 'multi-muscle group compensation to mask the fatigue characteristics' and 'long-time window heart rate analysis to be delayed inevitably', solve the calibration problem caused by individual variability, and are the technical problems to be solved in the field of the current wearable power assisting equipment and real-time fatigue monitoring. Disclosure of Invention The invention aims to overcome the defects in the prior art and provide a multi-muscle group fatigue detection algorithm based on the accumulation of the time sequence phase difference of the myocardial bi-modal signals. The technical scheme adopted for realizing the technical purpose of the invention is that the multi-muscle group fatigue detection algorithm based on the accumulation of the time sequence phase difference of the myocardial bi-modal signals comprises the following steps: step S1, acquiring surface electromyographic signal sequences and electrocardiosignal sequences of a plurality of target muscles, synchronously acquiring surface electromyographic signals and electroca