Signal processing of sEMG is currently the most prevalent method used to control active hand prostheses. For control purposess EMG is typically transformed into a feature space representation prior to presentation to a controller or classifier. In this study we compare three signal processing techniques for myoelectric control based on low level EMG contractions: mean-absolute-value (MAV), a Bayesian estimate of the EMGs ‘neural drive’, and sequentially updated real-time Kurtosis.
|Number of pages||1|
|Publication status||Published - 18 Aug 2017|
|Event||Myoelectric Controls Symposium 2017 - Fredericton, Canada|
Duration: 15 Aug 2017 → 18 Aug 2017
|Conference||Myoelectric Controls Symposium 2017|
|Abbreviated title||MEC 2017|
|Period||15/08/17 → 18/08/17|