DIGITAL TWIN OF A HAND EXOSKELETON FOR MODEL-BASED CONTROL AND FUNCTIONAL SAFETY IN REHABILITATION SCENARIOS

This thesis develops a comprehensive Digital Twin of a hand exoskeleton, integrating physics-based modeling, cascade control (admittance + PD with feedforward), three‑layer functional safety (enforcement, detection, governance), and observer‑based fault detection (Luenberger, generalized momentum, etc.). The closed‑chain mechanism is handled via a constraint formulation on URDF, reducing the dynamics to a scalar nonlinear ODE in the single actuated degree of freedom. The model includes passive finger response (Fung springs). Safety implements watchdogs, debounced escalation logic, and a four‑state state machine, with fault injection testing. The hardware (BLDC actuators, drives, amplifiers) is fully documented. The Digital Twin supports the entire development cycle from early design to safety verification prior to hardware‑in‑the‑loop integration.