Real-time classification of simultaneous hand and wrist motions using Artificial Neural Networks with variable threshold outputs

Paper in proceeding, 2013

Limb motions normally involve more than one degree of freedom combined in a coordinated manner. Although prosthetic hardware today could be combined for a highly motorized limb replacement, the control options available to amputees are so limited that this approach is rarely used. In this work, we introduce a classification strategy for the real-time simultaneous prediction of the individual movements present in natural motions. The real-time evaluation of this strategy based on a Multi-Layer Perceptron (MLP) with variable threshold outputs resulted in high motion completion rates. Moreover, the MLP alone showed higher offline accuracy than previously reported. This classifier was developed and evaluated in BioPatRec, an open source framework for advanced prosthetic control strategies based in pattern recognition algorithms. The source code and the data obtained in this study are freely available to be used for further algorithms development and benchmarking.