Acquisition and identification of myoelectric signals from the right forearm (AIM-SIM)

• Main Authors: Cobarrubias, Ann Kristine B., Doria, Mark Noel M., Jao, Dexter U., Quinzon, Marvin S.
• Format: text
• Language: English
• Published: Animo Repository 2008
• Subjects: Myoelectric prosthesis; Prosthesis; Amputees; Artificial arms > Power supply; Human engineering
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/14602
• Institution: De La Salle University

The acquisition and the processing of myoelectric signals are commonly used in biomedical engineering in conjunction with the development of prostheses, but it may also be applied in the development of machine interfaces that can replace controllers like the mouse or joysticks. A machine interface, which makes use of myoelectric signals as control input, addresses the hand amputees’ incapacity to control devices. It solves the problem without requiring, as a prerequisite, the complex surgical procedures that are involved in the development of prostheses. Using surface mount electrodes as sensors, the Acquisition and Identification of Myoelectric Signals from the Right Forearm (AIM-SIM) acquires myoelectric signals from the right forearm while the subject attempts to move particular muscles that are associated with the movement of specific fingers, and then sends the acquired signals to a computer for analysis. Two different methods for extracting the signals’ features are tested; the first technique uses Discrete Wavelet Transform (DWT) and Principal Component Analysis (PCA); the second technique uses Autoregressive Model (AR). The extracted features are inputted to a Multilayer Perceptron (MLP) that identifies which fingers the signals correspond to. The tests show that the first feature extraction method (DWT and PCA) is not suitable for the signals of interest. By using the outputs of the second method (AR) as inputs to the MLP, the resulting accuracy rating is 19.2% when BIOPAC MP35 is used; when the AIM-SIM acquisition device is used, the accuracy rating is 60%. The computer displays the result of the analysis in the form of a graphical hand. The output of AIM-SIM can be used to help increase the productivity of hand amputees by enabling them to control devices, such as computers, through the myoelectric signals in their forearms.