Design and Integration of an Event-Driven Rehabilitation System for Functional Electrical Stimulation

Spinal cord injury (SCI), stroke and multiple sclerosis are some of the main causes that lead people to paralysis, worsening their daily-life quality. In such scenario, physiotherapeutic procedures require novel techniques in order to improve effectiveness and suitability of rehabilitative applications. A proper response to this challenge can be the Functional Electrical Stimulation (FES), a technique born in the 1960s which is continuously evolving to satisfy medical requirements. The principle behind FES is to induce muscle contraction by applying low-energy pulses to the motor nerves of the muscle of interest, stimulating their activation through the use of non-invasive electrodes. Therefore, the development of a system able to enhance the benefits of FES application covers a central role in clinical and medical research programs. In this thesis project, a versatile multi-platform application for FES real-time control has been designed. Integrating previous and parallel works, the proposed software architecture allows a more efficient and embedded link between the FES stimulator and a surface ElectroMyoGraphy (sEMG) acquisition system. The sEMG is a non-invasive technique to acquire the electrical signals generated by skeletal muscles during contraction, from which it is possible to extract several parameters indicating the behavior of a muscle, such as its activity or its fatigue. However, the amount of information contained within these parameters and the processing complexity are constraints for low-power and real-time performance. For this reason, the proposed system exploits an event-driven approach applying the Average Threshold Crossing (ATC) to the sEMG signal. This technique works by generating an event whenever the sEMG signal exceeds a defined threshold. The average number of TC events inside an observation window has been proved to be an indicator of muscle activity; thus, this is the only value provided by the acquisition system. The central task of the application is to modulate the FES parameters (e.g., pulse amplitude, pulse width), according to the ATC values, and send them to the electrical stimulator. The application has been developed in Python with the ambition to satisfy the following requirements: modularity, scalability, extendibility, reliability. In order to accomplish this task, Object-Oriented Programming (OOP) has been exploited. The application is designed on three layers: at the bottom layer, two classes define the sEMG acquisition system and the electrical stimulator; at the middle layer, another class integrates the two previous ones, putting them in communication with each other; finally, at the top layer the Graphical User Interface (GUI) is defined. Implementing the multithreading paradigm, safety, usability and timing performance are achieved fulfilling the constraints of real-time application.