Adaptive Control in Neurorehabilitation

Abstract

Control theory, focusing on controlling physical systems, has been studied for long. In order to face diverse situations different control strategies have been defined that rely in aspects like predictability, knowledge of system behavior, parameters estimation and so. The field of rehabilitation, and specially neuroreabilitation, poses new questions and challenges as the system to be controlled is a robot that cooperates with a human, which may act in an unpredictable and random way. The human-robot physical contact and the need to assure human integrity constitute additional conditioning factors. This paper describes a control strategy that is an extension of the classical adaptive control approach, oriented to include a human in the loop, which interacts with the robot through neurological signals. An additional fact is the still very poor interpretation of neurological input commands which results in uncertain and ambiguous inputs to the rehabilitation robot controller.

This work is supported by the project HYPER, CSD2009 00067 Consolider Ingenio 2010.