Abstract  This paper describes a novel form of robotic therapy for the upper extremity in chronic stroke. Based on previous results, we hypothesized that a training task that encourages subjects to consciously guide endpoint forces generated by the hemiparetic arm will result in significant gains in functional ability of the arm, superior to more conventional methods of therapy. In addition, since stroke survivors present with varying degrees of arm movement ability, we developed an adaptive algorithm that tailors the amount of assistance provided in completing the guided force training task. The algorithm adapts a coefficient for velocity- dependent assistance based on measured movement speed, on a trial-to-trial basis. The training algorithm has been implemented with a simple linear robotic device called the ARM Guide. One participant completed a two month training program with the adaptive algorithm, resulting in significant improvements in the performance of functional tasks. Keywords  Robotics, Stroke, Rehabilitation, Adaptive control I. INTRODUCTION Robotic devices have been investigated as tools in upper extremity rehabilitation for chronic stroke survivors [1-4]. The first device used as a therapeutic tool, the MIT- MANUS, demonstrated that arm function in stroke survivors can benefit from interacting with an actuated planar device in the subacute stages of recovery [5]. Subsequently, two devices, MIME [4] and ARM Guide [3], expanded the investigations of therapeutic applications of robots into the chronic stroke population. These two studies verified that repetitive interaction with a mechanical device can result in improved performance of functional tasks. However, the outcomes of these two studies differed in that the reach extents of participants after eight weeks of training was different between the two devices. Participants in the MIME study were able to reach farther towards a target during unsupported and unguided movement, while users of the ARM Guide showed no change. After examining the differences between the methods in the two studies, we hypothesized that the fact that subjects consciously guided endpoint forces towards the reaching target with MIME may have been most responsible for the difference [6]. We decided, therefore, to test this guided force method with the ARM Guide One difficulty that arises in designing such a training paradigm in a stroke population is that participants exhibit a wide range of arm impairment levels. Because of this, some subjects are able to move through a large range of motion at a high velocity while others have severe range and velocity limitations. Adapting the level of assistance on a patient- specific basis is desirable because it would allow each patient to achieve as normal-as-possible movement, but with the least amount of assistance possible. This paper reports the development of a robotic movement training algorithm that couples guided force training and adaptive assistance. II. METHODOLOGY The new training method was implemented with the ARM Guide (Fig. 1). The device consists of a servo motor (M) controlling the position of the subjects arm (A), which is coupled to the device through a handpiece (H), along a linear track in the reaching direction. Two other degrees of freedom about the yaw and elevation axes (arrows) allow subjects to practice reaching movements to different areas of the workspace. A single trial, however, is performed with these two axes locked and movement only in the reaching direction. A six-axis load cell (F) reports the forces and torques at the interface between the subject and the device. The device is statically counterbalanced (C) so that it does not gravitationally load the arm. For a more detailed description, see [3]. Fig. 1. Diagram of the ARM Guide Adaptive Assistance for Guided Force Training in Chronic Stroke L. E. Kahn 1,2 , W. Z. Rymer 1,2,3 , D. J. Reinkensmeyer 2,4 1 Department of Biomedical Engineering, Northwestern University, IL, USA 2 Sensory Motor Performance Program, Rehabilitation Institute of Chicago, IL, USA 3 Department of Phys. Med. and Rehabilitation, Northwestern University Feinberg School of Medicine, Chicago, IL, USA 4 Department of Mechanical and Aerospace Engineering, University of California - Irvine, CA, USA F H A M C