Robotic arm controller using muscular contraction classification based on independent component analysis

Abstract

We develop a multi-channel electromyogram acquisition system base on the programmable system on chip (PSOC) microcontroller to control robotic arm. The array of 4x4 surface electrodes which invents from the low-cost EKG electrodes is used as the input sensor. B-spline interpolation technique has been utilized to map the EMG signal on the muscle surface. The topological mapping of the EMG is then analyzed to classify the pattern of muscle contraction. The proposed system was successfully demonstrated to record EMG data and its surface mapping. The muscular-contraction classification of mapping is then applied using independent component analysis. The classification result is then applied to control the movement of the robotic arm.