AUTHOR=Topini Alberto , Sansom William , Secciani Nicola , Bartalucci Lorenzo , Ridolfi Alessandro , Allotta Benedetto 
  
TITLE=Variable Admittance Control of a Hand Exoskeleton for Virtual Reality-Based Rehabilitation Tasks
  
JOURNAL=Frontiers in Neurorobotics
  
VOLUME=Volume 15 - 2021
  
YEAR=2022
  
URL=https://www.frontiersin.org/journals/neurorobotics/articles/10.3389/fnbot.2021.789743
  
DOI=10.3389/fnbot.2021.789743
  
ISSN=1662-5218
  
ABSTRACT=Robot-based rehabilitation is consolidating 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. 
Whenever the patient grasps a virtual object, the HES provides the user with a force feedback sensation.
The development process has been carried out at the Department of Industrial Engineering of the University of Florence (UNIFI DIEF) within the BMIFOCUS research project.
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 deeply investigated. 
The statements above will be thoroughly supported by a quantitative analysis of the achieved experimental results.