Channel selection for simultaneous myoelectric prosthesis control

Abstract

To develop a clinic ally available prosthesis base d on electromyography (EM G) signals, t he n umber of r ecording electrodes should be as small a s p ossible. In t his stud y, we investigate the po ssibility of th e least absolute s hrinkage and selection operator (LASSO) for finding electrode subsets suitable for regression based myoelectric prosthesis control. EMG signals were re corded using 192 electrodes while ten s ubjects were performing tw o degree-of-freedom (DoF ) wrist movem ents. Among the w hole c hannels, we s elected subsets consisting of 96, 64, 48, 32, 24, 16, 12, and 8 ele ctrodes, respectively, u sing the LASSO metho d. As a baseline method, electrode subset s ha ving the same numbers of electrodes w ere a rbitrary s elected w ith regular spacing (uniform selection method). The performance of decoding the movements was estimated using the r-square value. The e lectrode su bsets s elected b y the LAS SO met hod g enerally outperformed those chosen by the arbitrary selection method. In particular, th e performance of th e LA SSO method was significantly higher th an that of the a rbitrary selection me thod when using the subsets of 8 e lectrodes. From the analysis results, we could confir m that the LASSO method can b e us ed to s elect reasonable el ectrode subsets for reg ression b ased m yoelectric prosthesis control.Keywords-electromyography ( EMG); my oelectric control; prosthetic hand; regression; least absolute shrinkage and selection operator (LASSO)

Proportional estimation of finger movements from high-density surface electromyography

“…Regular electrode grids are convenient for practical implementation and allow simple mounting, since they can be realized as extensible uniform bracelets that are simply wrapped around the forearm [ 56 ]. Future investigations will further address the minimization of the number of channels, determining acceptable electrode locations and optimizing electrode-recording configurations, using the established methods for feature reduction [ 57 – 59 ]. Importantly, the results of the present study (offline tests) demonstrate that the number of electrodes can be decreased substantially (from 192 to 10) without significantly compromising the performance.…”

A hybrid auricular control system: direct, simultaneous, and proportional myoelectric control of two degrees of freedom in prosthetic hands

“…Examples are parallel pattern recognition [22], or musculoskeletal model based approaches [23]. More recently, regression-based approaches have been introduced to allow for direct and simultaneous control of multiple DOF [8,[24][25][26][27][28][29][30][31]. Due to the capability of the user to adapt for shortcomings of the control in the regression based approach [32], many problems of classification based systems such as small electrode displacements [33,34], variation of the arm [20,[35][36][37] or wrist [38] position, changes in skin conditions [19] and transfer between training and application [39] seem to be less problematic.…”