Reciprocal Kinematic Control: Using Human-Robot Dual Adaptation to Control Upper Limb Assistive Devices (bibtex)
by , , , ,
Bibtex Entry:
  author    = {Legrand, M and de Montalivet, E and Richer, F and Jarrassé, N and Morel, G},
  title     = {Reciprocal Kinematic Control: Using Human-Robot Dual Adaptation to Control Upper Limb Assistive Devices},
  booktitle = {Proceedings of the Hamlyn Symposium on Medical Robotics},
  year      = {2019},
  pages     = {xx},
  abstract  = {Upper limb (UL) assistive robots, such as exoskeletons, prostheses or supernumerary limbs, can rarely be fully autonomous devices. Indeed, it is generally not possible to use pre-defined patterns of motions because of the great diversity of tasks and the variety of UL movement strategies to achieve any of them. Control has to be provided to the users. For that purpose, the most widespread solutions to obtain user’s motor intention use physiological signals (electromyograms or electroencephalograms e.g.), distal functional joints (for instance, head or foot motions control the end-effector position and/or orientation) or inter-joint synergies models[5]. Despite interesting results, they all still have important limitations: the first two are neither natural nor intuitive and suffer from robustness issues, the third one does not allow very versatile devices. To tackle these issues, we propose a new control approach, together with a new paradigm, that uses the motion strategies naturally developed by the Central Nervous System (CNS). When a limb mobility is reduced, or when an assistive device does not work properly, CNS compensates and takes advantage of motor redundancy of the body: it calls other joints to still perform the desired gesture. Typical compensatory joints for UL movements are the trunk and the scapula. Our concept is to servo the robot to these body compensations. The only task of the latter is to make its user come back to a comfortable posture, and this indirectly leads to the realisation of the intended motion. The reciprocal adaptation between human and robot allows both to reduce the body compensations and perform UL movements with the assistive device. We validated a proof of concept of this paradigm on ten healthy subjects who executed a path-tracking task with an elbow exoskeleton.
  category  = {ACTIS},
  crac      = {n},
  hal       = {y},
Powered by bibtexbrowser