Abstract
Robot-aided neuro-rehabilitation is a potential solution to help resolve the increasing demand for post-stroke therapy. In this paper, we propose an assistive-resistive approach to promote active patient participation during robotic therapy. In order to automatically determine the most appropriate level of task difficulty, we recommend a novel approach for real-time adaptation of the robotic assistance according to the instantaneous patient's participation and performance. First, we estimated the dynamic contribution of the patient from robot's torque and kinematic data. Second, we adapted the stiffness parameter of the robot's impedance control according to the purported patient's participation. And third, we adapted the robotic assistance according to his/her performance while playing a serious game. We present the results of a feasibility study in which we evaluated our methodology with four persons with impairments due to stroke in a single session using the ankle robot. Our proposed metric for performance and participation has potential implications for real-time assessment during robot-assisted therapy. The proposed adaptive strategy, when compared to a fixed-stiffness approach, allows more kinematic variability in patient-led movements without compromising the overall performance during therapy.
Funding Information
  • Coordenao de Aperfeioamento de Pessoal de Nvel Superior do Brasil (CAPES) - Finance Code 001 (PDSE-88881.133763/2016-01)
  • São Paulo Research Foundation (2011/10369-6 and 2013/14756-0)
  • MIT-Brazil Program

This publication has 33 references indexed in Scilit: