Iterative Learning Synchronization of Robotic Rehabilitation Tasks

Abstract

Functional and task-oriented robotic rehabilitation training can be improved by the interaction of the patient with real-world objects. If these objects are not directly coupled to the robot, a synchronization of robot and object is required, comparable to a calibration between the end effector of an industrial robot and a work piece. Such synchronization prevents from safety hazards and unwanted tension forces at the man-machine interface between robot and patient or patient and object. In this paper, we propose a combination of an iterative learning synchronization (ILS) and an immediate error correction (IEC) technique to automate such synchronization processes. The feasibility of these methods is demonstrated by using them to automatically synchronize the gait rehabilitation robot Lokomat with a treadmilL.