2006 基于非线性尺度小波变换的表面肌电信号的分类_胡晓.pdfVIP

J Biomed Eng 2006; 23( 6) * 1, 2 1$ 1 1 1 胡　晓 　王志中 　任小梅　颜志国　王　刚 1( 上海交通大学 生物医学工程系, 上海 200030) 2( 广州大学 信息与机电工程学院, 广州510006) (Surface EMG sEMG, ) (Wavelet transform with nonlinear scale, WT ) WT , sEMG -, WT sEMG (30 30sEM G ) ( ) , ,,, BP : , WT , BP Classification of Surface EMG Signal Based on Wavelet Transform with Nonlinear Scale 1,2 1 1 1 1 Hu Xiao Wang hizhong Ren Xiaomei Yan higuo Wang Gang 1(D ep art ment of B iomed ical Eng ineer i ng , S hanghai J iao to ng Univ ers ity , S hang hai200030, China) 2(I nf o rmati on, M achinery and E lec tro nics Co ll eg e, G uangz hou Univ er si ty , Guang z ho u 510006, Chi na) AbstractSurface EMG (sEMG) signal is a complex nonlinear, non-stationary signal. In this paper, wavelet transform with nonlinear scale ( WT ) is introduced. Due to the gradual shortening of its time-resolution, WT is good at extracting the precise time-frequency information from sEMG signal. First, every sEMG signal ( 30 sets are for forearm supination and 30 sets are for forearm pronation is transformed into intensity distribution time) ( - frequency distribution) by WT. And then the feature vector is determined from the characteristic roots which are obtained from the intensity distribution by principle component analysis. At last, the two patterns of sEM G signals are identified by BP neural network T he results show that the accurate classification rate is higher gained. - . , by WT than by two conventional time frequency distributions At the same time the calculating complexity of neural network is decreased greatly.