Hybrid control of a robotic arm using EEG and EMG signals

Abstract

The human arms and legs play a very important role in performing daily life activities. The arm physiology is a very complicated subject of study. Amputee persons face difficulties due to the lack of their legs or arms but artificial limb can make their lives easier. The artificial limb may be upper limb or lower limb based on bio-medical signals acquired from the subject. In this context, the bio-medical signal is acquired, pre-processed, suitable features are extracted and classified in order to identify the corresponding motion and generate the desired control signal for controlling the robotic arm. The worth mentioning point here is the classification accuracy because higher the accuracy, more accurate is the control of the robotic arm. newlineThe surface electromyography (SEMG) signal is a biomedical signal used in various engineering and medical applications. The SEMG signal classification plays a key role for designing assistive devices for amputees and older age persons. Selection of suitable features plays a pivotal role in Electromyography pattern recognition (EMG-PR) based system designing. Time-domain features are widely used in EMG-PR based applications and indicated improved proficiency in the development of rehabilitation robotics, even though, the performance of existing features is not satisfactory. newlineIn the present work, the investigator proposed four novel time-domain features obtained by using first-order differentiation of original SEMG signals along with newlinethe ten time-frequency domain features based on discrete wavelet transform. The performance of the proposed features was compared with domain and frequency domain features using different classification techniques. The data acquisition and pre-processing stage were carried out followed by the feature extraction process for better classification results. Experimental results demonstrate that the proposed features extracted by using first-order differentiation of SEMG signals achieved better classification accuracy with SVM classifier. newline newline