Variable Admittance Control of a Hand Exoskeleton for Virtual Reality-Based Rehabilitation Tasks
Alberto Topini,William Sansom,Nicola Secciani,Lorenzo Bartalucci,Alessandro Ridolfi,Benedetto Allotta
DOI: https://doi.org/10.3389/fnbot.2021.789743
IF: 3.493
2022-01-12
Frontiers in Neurorobotics
Abstract:Robot-based rehabilitation is consolidated as a viable and efficient practice to speed up and improve the recovery of lost functions. Several studies highlight that patients are encouraged to undergo their therapies and feel more involved in the process when collaborating with a user-friendly robotic environment. Object manipulation is a crucial element of hand rehabilitation treatments; however, as a standalone process may result in being repetitive and unstimulating in the long run. In this view, robotic devices, like hand exoskeletons, do arise as an excellent tool to boost both therapy's outcome and patient participation, especially when paired with the advantages offered by interacting with virtual reality (VR). Indeed, virtual environments can simulate real-life manipulation tasks and real-time assign a score to the patient's performance, thus providing challenging exercises while promoting training with a reward-based system. Besides, they can be easily reconfigured to match the patient's needs by manipulating exercise intensity, e.g., Assistance-As-Needed (AAN) and the required tasks. Modern VR can also render interaction forces when paired to wearable devices to give the user some sort of proprioceptive force or tactile feedback. Motivated by these considerations, a Hand Exoskeleton System (HES) has been designed to be interfaced with a variable admittance control to achieve VR-based rehabilitation tasks. The exoskeleton assists the patient's movements according to force feedback and following a reference value calculated inside the VR. Whenever the patient grasps a virtual object, the HES provides the user with a force feedback sensation. In this paper, the virtual environment, developed within the Webots framework and rendering a HES digital-twin mapping and mimicking the actual HES motion, will be described in detail. Furthermore, the admittance control strategy, which continuously varies the control parameters to best render the force sensation and adapt to the user's motion intentions, will be investigated. The proposed approach has been tested on a single subject in the framework of a pilot study.
robotics,computer science, artificial intelligence,neurosciences