Development of an interface for real-time myoelectric control of external devices

The loss of upper limb functionality is a debilitating condition affecting many individuals. While the development of external devices has achieved remarkable capabilities in replicat- ing upper limb functions, there remains a critical need for effective control strategies. The objective of this experimental Master’s Thesis is to develop and validate a software appli- cation for the real-time translation of the electrical activity of forearm muscles into hand movements. Additionally, the solution provides a tool for visualizing kinematics output. The software application was developed in Simulink as an extension of the MEACS system, the EMG amplifier developed by LISiN - Politecnico di Torino. The application receives EMG input through TCP/IP communication, processes the signal, translates it into hand kinematic state through a neural network, and displays the output as a 3D plot. The appli- cation also includes a Graphical User Interface for initiating the TCP/IP communication, adjusting signal processing parameters, and saving predicted angles. The primary challenge of this work is to ensure real-time performance, providing visual feedback of the predicted hand gestures during their execution. The experimental results demonstrate the feasibility and potential of this approach for real-time implementation. However, it is crucial to acknowledge that the performance of the interface, specifically the accuracy of hand kinematics prediction, depends on the neural network employed. Despite these challenges, the proposed interface could represent a significant step towards addressing the critical need for effective control strategies in the field of upper limb prosthet- ics. In the future, this interface could be integrated with upper limb prostheses, enabling individuals to control their artificial limbs based on forearm muscular activity, potentially mirroring their intentions and improving their quality of life. This MSc project stems from the collaboration between LISiN at Politecnico di Torino and the Nuffield Department of Surgical Sciences at the University of Oxford. It is intended as a contribution to the doctoral research led by Giovanni Rolandino at the University of Oxford.