The eye-gaze-based approach to the control of a 5-degree-of-freedom upper extremity active exoskeleton

Stroke is a leading cause of adult disabilities, resulting in conditions like paresis and hemiparesis, significantly impeding daily activities resulting in aggreviation of their indepenence and quality of life. To aid their activities of daily living execution and enhance recovery process, a novel approach using eye-gaze tracking is proposed to control exoskeleton robots, offering an inclusive and accessible method for robotised rehabilitation services and general motion support. Most current solutions incorporate expensive eye-tracking devices, while the included method utilises a low-cost video camera, resulting in broader accessibility. An efficient algorithm for eye-tracking and blinking detection was introduced along with the design of the graphical user interface (GUI) suited to control the 5-DOF upper extremity \textit{ExoReha} powered rehabilitation exoskeleton and implemented the calibration procedure adjusting the system for the wide range of conditions. It enables users to directly control the position of three powered joints, set the velocity of their motion or replay predefined trajectories using eye movements and blinking. Five participants took part in the study to examine the applicability and usability of the designed system. The result regarding eye-tracking accuracy was of 69.69\%. The observed outcomes of the GUI functionalities triggering mechanism based on the blinking recognition highlighted the need for further development.