A Motor Point Identification Technique Based on Dempster Shafer Theory

The objective of this work is to identify the motor point location from the obtained sEMG signals using Dempster Shafer theory (DST). The proposed technique is applied on data obtained from a male test subject. In particular, the sEMG signals and its corresponding skeletal muscle force signals from the Flexor Digitorum Superficialis are acquired at a sampling rate of 2000 Hz using a Delsys Bangnoli- 16 EMG system. The acquired sEMG signals are rectified and filtered using a Discrete Wavelet Transforms (DWT) with a Daubechies 44 mother wavelet. For the system identification, an Output Error (OE) model structure is assumed to obtain the dynamic relation between the sEMG signal and the corresponding finger force signals. Subsequently, model based probabilities and fuzzy inference based probabilities are obtained for discrete sensor locations of a sEMG sensor array. Considering these evidences, a DST based motor point location identification method is proposed. The results based on one subject show the potential of the proposed theory and approach for affectively identifying motor point locations using an array sEMG sensor.