The Impact of Different Exercise Visualization Approaches on Motion Accuracy During Exoskeleton-Aided Physiotherapy of an Upper Extremity – Initial Investigation

Abstract

Rehabilitation robots support specialists physically and automate tasks. However, exercising with them requires a patient to understand the intended therapy movements. This paper explores the impact of visual interfaces on the accuracy of robot-aided physiotherapy motions. The experimental study compares three visualisation types: 2D with bar sliders, 3D with phantoms, and immersive 3D in virtual reality with phantoms. Evaluation criteria include motion accuracy, muscle loads based on the registered accelerations, and user feedback. The experiments involve trials with a passive exoskeleton setup and three visual interfaces. Participants follow the presented daily life-like movements while the sensors measure their real joint rotations. The trajectories are compared based on six dynamic-related accuracy indicators. Results show that phantom-based interfaces significantly improve motion accuracy compared to 2D, with participants favouring these methods. However, no significant superiority of the virtual reality-based method was observed. The study highlights the importance of choosing appropriate visualisations for effective robot-aided physiotherapy. The observed trends can be used to design user interfaces for exoskeletons but also end-effector rehabilitation devices.