Robust, Limb Position Invariant, EMG-Based Finger Movement Recognition.

It is generally known that when changing the limb position or angle the performance of the EMG-based pattern recognition system greatly degrades. In such a case, many recent articles employed accelerometers with EMG sensors to account for the different limb positions. We prove here that by using moments invariants then the effect of the different limb position and/or angles can be greatly reduced with robust performance. More details after the paper publication.