AS-ULDAR: a ROS 2-based collaborative robotic platform for upper-limb rehabilitation — systems integration and pilot usability evaluation in a clinical environment

Robotic rehabilitation can increase the intensity and personalization of upper-limb therapy, but its clinical translation is hindered by the difficulty of integrating individually validated technical components into a single clinician-ready platform. This paper presents a platform which is intended to be an Auto-calibrating System for Upper Limb Disability Assessment and Rehabilitation (AS-ULDAR), a ROS 2-based platform built around a Franka Emika Panda robot that unifies motor/cognitive assessment, sensorless force estimation, and Cartesian and Human Joint-based Trajectory Tracking within a clinician-oriented GUI. While these functionalities were individually validated in prior work, the present contribution lies in their systems-level integration and in identifying the operational barriers that emerge when deploying such a platform in a real clinical workflow, targeting adult patients with chronic neurological conditions (stroke, Parkinson’s, ALS, MCI) and mild-to-moderate upper-limb impairment. The platform was evaluated through a formative usability study: the standalone GUI by five naïve users, and the complete system by seven physiotherapists performing the full clinical workflow (no patients were involved; clinical efficacy is outside this scope). The system was technically stable throughout. Usability (SUS) was high for the GUI (87.5 ± 1.8) but markedly lower and more heterogeneous for the complete platform (54 ± 16), with feedback pointing to robot setup and workflow operation—not interface design—as the main friction points for clinical staff. These results show that integrating multiple rehabilitation functionalities is technically feasible, but that usability must be treated as a system-level property, and they provide concrete operational lessons for the clinical translation of cobot-based rehabilitation platforms.