Digital Twin and Resistive Control Strategy for the FloatEVO Upper Limb Rehabilitative Exoskeleton

This work was funded by the Istituto Nazionale per l’Assicurazione contro gli Infortuni sul Lavoro (INAIL) under grant agreements "PR23-RR-P2 – ClinicEXO". Stroke is a cerebrovascular accident caused by the interruption of blood supply to the brain, either due to blocked artery (ischemic) or bleeding (haemorrhagic). It is a major cause of long-term disability worldwide, often resulting in hemiparesis or upper limb motor impairment, which significantly constrains post-stroke patients’ ability in performing activities of daily living. To address these challenges, rehabilitative exoskeletons are designed for supporting the recovery process. FloatEVO, the reworked version of Float upper limb exoskeleton, has 7 degrees-of-freedom allowing for the rehabilitation of scapular, glenohumeral, and elbow flexion movements. It is designed for orthopaedic (trauma or reverse prosthesis), as well as neurological (stroke or trauma) patients. The exoskeleton enables intensive recovery sessions, by precisely guiding patients through rehabilitation exercises, a higher number of repetitions and adjustable controls suitable for each user, making the rehabilitation process truly patient-centred. The first part of this thesis focuses on the implementation of new rehabilitation exercises in FloatEVO, suggested by clinicians. At the beginning, interesting novel exercises were studied on healthy subjects (without the exoskeleton) using Vicon motion capture system. Data was collected to obtain the joint trajectories of the human upper limb during horizontal abduction exercises with different planes of elevation at 30°, 45°, 60°, 90°, and 120°, as well as abductions with elevation angles of 0°, 30°, 60°, and 90°, and then analysed in Matlab. The horizontal abductions (excluding the one at 45°) were simulated in Unity environment after constructing a virtual model of FloatEVO. Next, those trajectories were coherently recorded with the exoskeleton and the corresponding log files were extracted to study the kinematic relationship among the joint space trajectories, specific to each exercise. The physiological movement represented by these trajectories (other than the one at 120°), was validated through surface electromyography (sEMG) with the real exoskeleton and further analysed in Cartesian space to ensure that the exoskeleton’s movements fell within the range of natural human motion. In the second part of the thesis, FloatEVO’s control system was updated to introduce a resistive control, apart from the assistive one. Besides a constant resistive torque, a Gaussian resistance profile was developed. This unconventional approach does not constitute real haptic feedback, however it has the potential to enhance the rehabilitation experience by delivering motivating and challenging feedback, which is especially advantageous for patients in the later stages of recovery. Future developments may include integrating visual feedback into augmented reality systems such as HoloLens, perhaps incorporating interactive elements (e.g., virtual obstacles like climbing a hill) to further engage and stimulate patients throughout their rehabilitation process. To evaluate the effectiveness of these new control strategies, sEMG data were collected from participants performing horizontal abductions at 30°, 60°, and 90° under varying assistance and resistance levels. These data provided insights into muscular activation across different control modes, assessing the performance and impact of the novel resistive control strategies.