Research On The Training Method Of Health-related Physical Fitness Improvement Based On The Analysis Of Muscle Fatigue Characteristics

Health-related physical fitness mainly includes four elements:cardiorespiratory endurance,muscular fitness,body composition and flexibility.Health-related physical fitness is the basic physical quality that characterizes human health status and disease risk.Improving human health-related physical fitness can effectively reduce the risk of cardiovascular and cerebrovascular diseases and metabolic syndrome.It is a hot spot in sports medicine and health promotion technology research.Ordinary people lack professional and scientific guidance,and it is difficult to improve their health-related physical fitness through exercise.At the same time,spontaneous and disordered exercise training is inefficient and prone to sports injuries.A safe,efficient,and personalized closed-loop training mode is urgently needed.In view of the above problems,this paper conducts a research on the training method of personalized health-related physical fitness improvement based on the analysis of muscle fatigue characteristics.Including:carried out surface electromyographic(sEMG)signal time-frequency domain synergistic analysis and sEMG signal denoising research,extract the target muscle sEMG signal features,build a deep learning model framework for dynamic muscle fatigue recognition,design a personalized health-related physical fitness.improvement training method system,and use muscle fatigue as an indicator for training effect evaluation and process monitoring.The main research work is as follows:1.Cyclic resistance incremental workload test.In this paper,we designed two experimental schemes.Experiment A:Obtain the sEMG signal data of the muscles during training from the non-fatigue state to the fatigue state,which provides the data basis for building a dynamic muscle fatigue state recognition model.Experiment B:Obtain the sEMG signal data during the training to provide the data basis for the formulation of personalized training prescriptions.Experiments A and B both require subjects to perform cyclic resistance incremental load training on the cycle ergometer to exhaustion.The sEMG signals of the rectus femoris,vastus lateralis,vastus medialis and gastrocnemius during training were acquired using a wireless sEMG signal acquisition device.In addition,before Experiment A and after each test period,we used a muscle tester to test the maximal voluntary isometric contractions(MVICs)of the subjects’ lower extremity muscles.The obtained MVICs data were used to evaluate the state of muscle fatigue during the experiment.2.Research on the sEMG signal denoising method.The collected sEMG signal was analyzed by the time domain analysis methods and frequency domain analysis methods,and found that it contained motion artifacts,power frequency interference and Gaussian white noise.First,we select the band-pass filter with the best filtering effect to remove motion artifacts and invalid signals in the original sEMG signal.An adaptive notch filter was designed to simultaneously eliminate 50 Hz harmonic power interference while retaining the effective signal within the frequency band.Then,an improved threshold function denoising method based on optimal wavelet packet tree transform was proposed to remove Gaussian white noise in sEMG signals.Signal to noise ratio(SNR)and root-mean-square error(RMSE)were used to objectively evaluate denoising performance.Finally,we designed a comparison experiment between the improved threshold function denoising method proposed in this paper and the traditional wavelet threshold function denoising method,and the commonly used threshold function denoising method in recent years.The experimental results show that the SNR indicator of the sEMG signal after denoising with the improved threshold function proposed in this paper is 75.41%higher than that of other denoising methods on average,and the RMSE indicator is 39.69%lower than other denoising methods on average.It is concluded that the improved threshold function proposed in this paper has better performance in eliminating Gaussian white noise.3.Research on the recognition model of exercise muscle fatigue state.First,we propose a muscle activation state sEMG signal detection algorithm,which can accurately identify and automatically segment sEMG signal fragments in the muscle activation state in the continuous sEMG signal sequence.The dataset was constructed by extracting the root mean square(RMS),integral EMG(IEMG),median frequency(MF),mean power spectrum frequency(MPF),MPF/mean absolute value(MAV)of each signal segment.Then we designed the Attention-LSTM muscle fatigue state.recognition model based on the attention mechanism,the bidirectional long-short-term memory network and the stacked long-short-term memory network.The Attention-LSTM model proposed was compared with BFA-GSVCM model and PSO-SVM model,which are commonly used for muscle fatigue identification.The comparison results show that the recognition accuracy of the Attention-LSTM model is 2.83%higher than that of the BFA-GSVCM model and 4.89%higher than that of the PSO-SVM model.4.Research on the training method system of individualized health-related physical fitness improvement.To improve the three core indicators of cardiorespiratory endurance,muscle strength,and body composition in healthy physical fitness,follow the principles of FITT-VP exercise prescription,and build a training method system for improving health-related physical fitness.The training type and training termination conditions of the method system are circuit resistance training and muscle fatigue,respectively.The training intensity,number of training groups,and training frequency are formulated according to the subjects’ healthy physical fitness and sEMG fatigue threshold(sEMGFT).We designed experiments comparing individualized health-related physical fitness training method with high-intensity interval training(HIIT)method and high-intensity resistance training method.The experimental results show that the effectiveness of the individualized health-related physical fitness training method on cardiorespiratory endurance improvement is similar to that of HIIT and better than that of high-intensity resistance training.The effectiveness of the individualized health-related physical fitness training method on muscle strength improvement is similar to that of high-intensity resistance training and better than that of HIIT.The subjects’ body composition was significantly improved after training with the three training methods.The effectiveness and feasibility of the individualized health-related physical fitness training method system for improving people’s health-related physical fitness has been verified.In this paper,based on the research on the sEMG signal denoising method,muscle activation state sEMG signal detection method,and dynamic muscle fatigue recognition model,we design a health-related physical fitness training method system that can provide safe,scientific,and personalized training guidance.