Preliminary Evaluation of the Passive Rehabilitation Phase of the REACT System for Tendon Therapy

Tendon injuries, particularly in the hand, require early and controlled rehabilitation to restore joint mobility while preventing adhesion and minimizing the risk of complications, such as suture rupture or gap formation. Robotic tendon therapy systems offer a promising alternative through programmable and repeatable motion delivery, yet many current designs are hindered by bulk and mechanical complexity, limiting their practicality in clinical use. Shape memory alloy (SMA) based actuators represent a promising solution for tendon-driven exoskeletons owing to their lightweight construction, high force-to-weight ratio, and silent operation, features that are particularly advantageous for wearable and portable rehabilitation devices. However, the nonlinear, hysteretic, and thermally dependent nature of SMA behavior poses significant control challenges, necessitating detailed experimental characterization and modelling. This study presents the tendon-driven Rehabilitation Exoskeleton for Assisted Coordination of Tendon Therapy (REACT) system, incorporating experimentally derived force-temperature and current-temperature models of SMA units into a current-based position control framework for the passive rehabilitation phase. The controller incorporates these models into the forward dynamics of the finger to achieve accurate torque generation and trajectory tracking within the physiological range of motion. Simulation results indicate high-fidelity reference tracking, with a steady-state error of approximately -8 x 10(-3) rad, thereby confirming the feasibility of SMA-based actuation for safe, precise joint positioning in early-stage tendon rehabilitation.