Patient-Cooperative Control: Providing Safe Support without Restricting Movement

Abstract

Patient-cooperative behavior of a rehabilitation robot can be seen as a tight interplay of three control components: The first and most important is the intervention paradigm, which can for example be assistance, resistance, or error augmentation. The second and third are more related to the underlying properties of the robot: On the one hand the robot should be transparent in “free” movements, and on the other hand it should provide a safe training frame with appropriate virtual constraints for the movement. In this paper, control strategies to enhance transparency and to constrain movement with virtual tunnels are presented using the examples of the ARMin and the Lokomat, which are rehabilitation robots for upper and lower extremities, respectively. Differences and similarities in control of these robots are outlined in terms of the control strategies for transparency enhancement and movement constraints. The control concepts Generalized Elasticities and Path Control are described, which improve transparency in free movements inside an allowed spatial region, and which impose movement constraints to confine the user to this allowed region. Generalized Elastic Path Control unifies both control approaches within a single potential field, and preliminary results of this controller on the Lokomat are shown.