A Cost-Efficient Tele-rehabilitation Device for Training Distal Upper Limb Functions After Stroke

Robotic rehabilitation devices offer prospects in improving the therapy outcome in stroke patients. In particular the combination with tele-rehabilitation functionality may be beneficial to reduce cost, which is especially required for home-based rehabilitation. In this paper a device is presented that allows for exercising supination/pronation, dorsiflexion, and finger training. Its communication architecture follows a modular design approach. The Qt-based graphical UI can be executed on different operating systems and devices including the cost-effective Rasperry Pi single-board computer. Tele-rehabilitation functionality is implemented based on SSL-encrypted RESTful web services following a three-tier architecture. Cost is reduced by omitting expensive sensors. A torque sensor is replaced with current-based torque sensing, used for progress measurement and interactive exercises. The evaluation shows accurate results after compensating the static friction, justifying the omission of an additional torque sensor. Torque measurements during passive exercises showed higher and more asymmetric ratings for a stroke patient than for a healthy subject indicating that this measurement may be used as an estimator of spasticity.