基于运动疲劳下语音参数变化的研究

Research on changes of speech parameters based on exercise fatigue

运动疲劳的检测在运动科学界有着非常重要的意义,若没有科学的运动体系作为指导,运动过量对运动员的身体会造成严重伤害.基于语音的运动疲劳程度检测,研究了基于短时平均能量、短时平均幅度函数、短时过零率、短时自相关函数和梅尔倒谱系数5个语音特征参数在运动疲劳下变化的规律.实验结果显示,随着运动疲劳程度的加深,时域和频域语音参数的幅值均呈现不同程度的上升,其中语音信号中短时平均幅度升高得最明显,短时过零率最缓慢.2种不同程度运动之间,短时过零率的增长率最快,短时平均能量的增长率最缓慢.

The detection of exercise fatigue is of great significance in the sports science community.Without a scientific athletic system as a guide,excessive exercise can cause serious injury to athletes′bodies.Based on the detection of speech fatigue level,the changes of five speech characteristic parameters based on short-term average energy,short-term average amplitude function,short-time zero-crossing rate,short-term autocorrelation function,and Mel cepstrum coefficient were studied under sports fatigue law.The experimental results show that with the deepening of exercise fatigue,the amplitudes of time-domain and frequency-domain speech parameters have increased to varying degrees.Among them,the short-term average amplitude of speech signals increases most obviously,and the short-term zero-crossing rate is the slowest.Between two different degrees of exercise,the short-term zero-crossing rate has the fastest growth rate.The short-term average energy growth rate is the slowest.