A cable driven robotic system to train finger function after stroke

Abstract

This paper presents a novel robotic interface to train intrinsic finger movements. The mechanical design, based on a cable system interacting with the fingers, offers the possibility of adapting the interface to accommodate various hand sizes and finger orientation. A main feature of the device is a clutch system, which consists of five clutches, one for each finger, that can be switched to three different modes: (i) rest mode: the fingers are mechanically maintained at a fixed position, (ii) passive (from the view of the interface) mode: the finger is free to move along the path defined by the cable, and (iii) active mode: the force generated by the motor is applied to the finger. With this mechanism, it is possible to train hand muscle function using only one actuator. The interaction with the subject is measured by means of a position encoder and five force sensors located close to the output. We describe the human-oriented design of our underactuated robotic interface, based on measured biomechanics. We detail the redundant safety mechanisms, the actuation, sensing and control systems and report the performance and preliminary results obtained with this interface. ©2007 IEEE.