EMG Wrist-hand Motion Recognition System for Real-time Embedded Platform

Abstract

Electromyography (EMG) signal analysis is a popular method for controlling prosthetic and gesture control equipment. For portable systems, such as prosthetic limbs, real-time lowpower operation on embedded processors is critical, but to date there has been no record of how existing EMG analysis approaches support such deployments. This paper presents a novel approach to time-domain classification of multichannel EMG signals harnessed from randomly-placed sensors according to the wrist-hand movements which caused their occurrence. It shows how, by employing a very small set of time-domain features, Kernel Fisher discriminant feature projection and Radial Bias Function neural network classifiers, nine wrist-hand movements can be detected with accuracy exceeding 99% -surpassing the state-of-the-art on record. It also shows how, when deployed on ARM Cortex-A53, the processing time is not only sufficient to enable real-time processing but is also a factor 50 shorter than the leading time-frequency techniques on record.

Real-Time Hand Gesture Recognition Using Temporal Muscle Activation Maps of Multi-Channel Semg Signals

“…A number of studies have been conducted to accurately classify hand gestures from the pre-recorded sEMG signals [5,7,8]. However, only a few studies [9,10,11] have focused on real-time hand gesture recognition using sEMG signals from the forearm. Most methods [7,9,10,12] rely on binning of the sEMG signals, computing a set of features (mean absolute value, waveform length, etc.)…”

Development of Sign Language Motion Recognition System for Hearing-Impaired People Using Electromyography Signal

“…Even though the performance metrics of these studies cannot be compared directly due to different experiment settings, they are helpful for qualitative comparisons. Among them, the studies of [ 43 , 62 ] had the high accuracy more than 99%; however, the number of recognized gestures was less than 10, and [ 62 ] needed about 1 s to perform the recognition. On the other hand, although the study of [ 42 ] needed only 3 ms, the accuracy is 85.1%.…”

“…Approximately 95% of the signal power was below 400–500 Hz, which required the sampling rate to reach 1000 Hz in order to gather all the information according to the Nyquist sampling theory [ 38 , 39 , 40 ]. However, studies employing a low sampling rate device still obtained a decent accuracy with different approaches applied [ 41 , 42 , 43 ]. Some of other studies only carried the experiment on single participant and developed systems which had high accuracy [ 44 , 45 , 46 ].…”

EMG Biometric Systems Based on Different Wrist-Hand Movements

“…overlapping [50]. The non-linear polynomial Kernel Principal Component Analysis (KPCA) [51] and Kernel Fisher Discrimination (KFD) analysis [52] have been successively used for projecting multi-class features. Both non-linear projection schemes are considered and are combined with different feature classifiers.…”