Abstract:
In spite of the great advances in the
mechanical and electronic components
of prosthetic hands, they still lack the
high number of degrees of freedom
present in the real human hand. That is
due, not to technical deficiencies, but to
the much reduced amount of
independent control signals available
when using surface electromyography
(s-EMG) from the forearm stump or
other artificial sensors. Cross-talk
between adjacent muscles produces
interferences that bury the s-EMG of the
target muscle and reduce selectivity.
In a single case study, surface-EMG
signals from an able-bodied subject’s
forearm were recorded with a surface,
5x13-electrode array while the subject
performed eleven different isometric
contractions. In order to reduce the
cross-talk between s-EMG signals from
different muscles, we applied a blind
source separation (BSS) technique
called JADE.
Although the results are not fully
conclusive, they indicate that BSS
techniques could provide an important
reduction in s-EMG cross-talk and
hence BSS is able to increase the
selectivity of recordings for myoelectric
control.
