A pilot study for assessing NAO humanoid robot assistance in shoulder rehabilitation.

Purpose: This study aimed to explore the potential application of NAO in guiding patients through rehabilitative exercises using external audiovisual stimuli, focusing on temporospatial control in terms of range of motion (ROM), execution time and movement smoothness. Methods: This is a preliminary analysis involving ten healthy volunteers and two patients with shoulder musculoskeletal disorders. The protocol was developed in two phases (III and IV) with different ROM limits and including flexion-extension (FE), external-rotation (ER) and internal-rotation (IR) exercises, performed at two speeds and both with and without NAO assistance. Simultaneously, upper limb kinematics were assessed using a stereophotogrammetric system as a reference. Performance was evaluated by mean absolute error (MAE) for ROM and execution time, with smoothness assessed through Log Dimensionless Jerk analysis. Results: In phase III, results for volunteers showed ROM differences in FE and ER, while IR was unaffected by NAO presence. In phase IV, NAO assistance resulted in reduced MAE across nearly all exercises. Patients who only performed phase III exercises at lower speed stayed within ROM limits for all movements performed with NAO, except for ER. For all the participants, results showed a significant reduction in the time MAE when using NAO. Patients exhibit greater smoothness during FE performed with NAO. Conclusions: NAO showed potential in aiding patients with shoulder musculoskeletal disorders to replicate rehabilitation exercises, guiding both ROM and timing while influencing movement smoothness. NAO imitation could lead to improved rehabilitation outcomes and enhanced motor learning of motor skills, fostering greater adherence to prescribed therapy. Level of Evidence: Level V, diagnostic.