Improving Wrist Angle Recognition Accuracy under Different Load Conditions

The wrist angle estimation based on surface electromyography (sEMG) signals plays an important role in the sEMG application. This paper confirms that the accuracy of the wrist angle recognition decreases with the increase of the wrist load by the changes of the sEMG features in different loads. To address the above problem, this paper proposes a combined feature, integrating frequency-domain and time-domain features, to improve the recognition accuracy, which has been demonstrated by comparative experimental results.