Study On Detecting And Processing Techniques For Dual-Phase Motion Fatigue Information Induced By Neuromuscular Electrical Stimulation

Nowadays, the main limitation to daily use of neuromuscular electrical stimulation (NMES) is the muscle fatigue induced by NMES. As a kind of inevitable physiological phenomenon in rehabilitation training, muscle fatigue has a direct influence on the rehabilitation effect of stimulation. Study on how to detect and evaluate the muscle fatigue level and even overcome the negative effect of fatigue will have great methodology significance to neuromuscular system cognition, disable rehabilitation engineering and clinical physiotherapy evaluation. Muscle activity can be recorded by surface electromyography (sEMG) and, therefore, some feature parameters of sEMG can be used to characterize the muscle fatigue induced by NMES.In this study, the experiment platform and scheme were designed for testing the muscle fatigue of lower limbs during static and dynamic phases induced by NMES. The sEMG collected in the experiments were processed by two-point peak threshold detection algorithm to remove the stimulus artifact of NMES, empirical mode decomposition (EMD) method to remove the low-frequency drift and adaptive trap filter to remove the influence of 50Hz and other harmonic wave disturbances. By detecting the changes of knee joint angle, the information about muscle fatigue induced by NMES could be obtained in static and dynamic phases. sEMG features during static phase were extracted as median frequency, mean frequency and autoregressive (AR) model index. Continuous wavelet transform was applied to analyze the sEMG recorded in mobile phase. The wavelet coefficient was classified into two parts with high frequency band and low frequency band. Root mean square value was used to evaluate the amplitude of sEMG in different bands and analyze the trend of these parameters in the fatigue process.Relevant results showed that in the muscle fatigue process induced by NMES during static phase, median frequency, mean frequency and AR model index were all reliable parameters to characterize the muscle fatigue and, during dynamic phase, the wavelet coefficient amplitude of sEMG in lower frequency band after wavelet analysis could be used to evaluate the muscle fatigue level in mobile phase.In this thesis, the knee joint angle detection and sEMG spectrum analysis were implemented with the muscle fatigue experiments of lower limbs during static and dynamic phases induced by NMES to acquire the signal feature parameters for evaluating the muscle fatigue, especially the wavelet coefficient amplitudes in higher and lower frequency bands for estimating the myodynamia and muscle fatigue, respectively, which may provide a feasible method to accurately assess and feedback the muscle fatigue induced by NMES.