Abstract
The voluntary contracted EMG signal of spinal cord injured patients is very small because the information from central nervous system is not sufficiently transmitted to $\alpha$ motor neuron or muscle fiber. Therefore the acquisited EMG signal from needle or surface electrodes can not be identified obvious voluntary contraction pattern by muscle movement. In this paper we propose the extraction technique of voluntary muscle contraction and relaxation pattern from EMG signal of spinal cord injured patient whose EMG signal is composed of the linear sum of mo색 unit action potentials with two noise sources, additive noise assumed to be white Gaussian noise and high frequency discharge assumed to be not motor unit action potential but impulsive noise. In order to eliminate impulsive noise and additive noise from voluntary contracted EMG signal, we use the FatBear filter which is a nonarithmetic piecewise constant filter, and multiscale nonlinear wavelet denoising processing, respectively. The proposed technique is applied to the EMG signal acquisited from transverse myelitis patients to extract voluntary muscle contraction pattern.
본 논물에서는 척수 손상으로 인하여 암이 수축 및 이완시 미약한 근전도, 신호를 발생시키는 환자로부터 명확한 수축 및 이완 패턴을 추출하기 위한 신호 처리 기법을 제안하였다. 제안한 방법은 비선형 고정 필터의 일종인 FatBear 필터를 이용하여 거대 운동단위 활동전위로 의심되는 충격 잡음을 제거하고 웨이브렛 평면에서 비선형 멀티 스케일 필터링 기법을 이용하여 가산 잡음을 제거하는 것으로서 횡단성 척수염으로 인한 마미 증후근을 보이는 환자들에게 적용하여 명확한 수축 및 이완 패턴을 추출할 수 있었다.
