Electromyography-based Control of Prosthetic Arm for Transradial Amputees using Principal Component Analysis and Support Vector Machine Algorithms

A prosthetic limb based on two different algorithms such as the Principal Component Analysis (PCA) and the Support Vector Machine (SVM); also, an acquisition, filtering, and amplification circuit which would be utilized to identify the incoming signals based on the nature of the movement. The prosthetic limb aims to mimic basic hand movements such as the grasping and releasing motion, which aims to help transradial amputees with simple tasks. The method of acquisition to be used is of a non-invasive type of probe which would be positioned at specific locations in the bicep area and a cascaded filtering circuit with a specified frequency response desirable for electromyography (EMG) extraction. The PCA in conjunction with the SVM, through the Python programming language, processes the signal through coded program and supervised machine learning which would be used to make decisions based on muscle movement, and determine the correct output signal to be sent out for actuation. All signals acquired and filtered would be fed to the Raspberry Pi 3, which is the used microcomputer in the paper, and then used as input for the two different algorithms. The paper aims to provide further application on the machine learning algorithms used specifically in the field of prosthesis which using a microcomputer, would provide real time acquisition, processing, and prediction that leads to actuation.