Method of sequential classification for pre-grouping hand movements classes for fast labeling process of an electromyographic signals

摘要

The paper discusses the problem of the process of quick recognition and classification of electromyographic signals in order to control a robotic prosthetic hand. The work is based on the assumption that during the recording of the movement, the signal contains useful information from the initial acquisition of measurement data. They can be used for early grouping of movement classes into larger sets, giving preliminary information about the type of intentions of the movement and at the same time simplifying the labeling process at a later stage.