A biomimetic and multichannel FES implementation for lower body rehabilitation protocols

The hip and knee joints are among the most mobile joints in the body and play an important role in supporting body weight and in allowing daily activities such as walking and climbing stairs. Due to accidents or pathologies, like osteoarthritis, such joints can be damaged. In severe cases, where motor functions are limited, the patient must undergo an arthroplasty surgery, which must then be followed by a rehabilitation program for the recovery of lost muscle strength. The aim of this project is to define two rehabilitation protocols for patients with hip and knee prostheses that can be implemented by FES. To achieve this goal, an event-driven closed-loop system for FES real-time control, designed in previous works, was used. This system allows to acquire surface ElectroMyoGraphy (sEMG) signals from a subject (therapist) through wearable acquisition boards and translate them into stimulation patterns to provide to another subject (patient) through surface electrodes in real-time. However, to optimize the acquisition sensors circuit and power consumption, instead of using all the information carried by the sEMG signal, a feature is extracted through the Average Threshold Crossing (ATC) technique. The latter is based on generating an event whenever the sEMG signal exceeds a certain threshold. By averaging the number of TC events within a time window, the ATC value is computed. Each time the therapist performs a movement, variations of ATC parameters of each muscle involved are detected and translated into modulations of current intensity (I) to send to the patient. In this thesis, two improvements to the system software are proposed: the possibility of combining the individual muscles’ activations, involved in the execution of a single movement, to generate a single stimulation pattern and the possibility of providing the patient with an alternative stimulation mode, pulse width modulation (PW). Then, the procedure followed to outline every single movement of the two protocols is described: which muscles were considered and how the relative signals were combined to obtain the activation and stimulation profiles and how the surface electrodes were positioned on the patient's muscles. Each protocol was tested on 3 subjects and for each movement, the two types of modulation (I and PW) were compared in terms of movement execution, patient discomfort and muscle fatigue. These data were obtained through forms filled out by patients at the end of each test, in which questions were asked about the discomfort experienced during the stimulation phase, the possible onset of fatigue during the repetitions of a movement and the level of muscle fatigue perceived at the end of the session. The results show a preference of 50-80% for PW modulation in the case of movements that involve keeping the limb in a certain position for a few seconds, as it allows to generate a greater contraction force. For movements that include small muscles, a preference of 66-100% has almost always been expressed for I modulation since it allows the patient to perform the movement without too much discomfort. Finally, the relationship between fatigue perceived during 10 movements and the charge delivery to the patient in each repetition, which is calculated by adding the charges of each pulse included in that repetition (single pulse charge delivery = IxPW of the pulse), was studied. If fatigue is perceived from a certain repetition onward, this also marks the beginning of an increase in the charge delivery to the patient.