Abstract
Rehabilitation robots have become important tools in stroke rehabilitation. Compared to manual arm training, robot-supported training can be more intensive, of longer duration and more repetitive. Therefore, robots have the potential to improve the rehabilitation process in stroke patients. Whereas a majority of previous work in upper limb rehabilitation robotics has focused on end-effector-based robots, a shift towards exoskeleton robots is taking place because they offer a better guidance of the human arm, especially for movements with a large range of motion. However, the implementation of an exoskeleton device introduces the challenge of reproducing the motion of the human shoulder, which is one of the most complex joints of the body. Thus, this paper starts with describing a simplified model of the human shoulder. On the basis of that model, a new ergonomic shoulder actuation principle that provides motion of the humerus head is proposed, and its implementation in the ARMin III arm therapy robot is described. The focus lies on the mechanics and actuation principle. The ARMin III robot provides three actuated degrees of freedom for the shoulder and one for the elbow joint. An additional module provides actuated lower arm pro/supination and wrist flexion/extension. Five ARMin III devices have been manufactured and they are currently undergoing clinical evaluation in hospitals in Switzerland and in the United States.
Acknowledgements
This work was supported in part by the ETH Foundation, the Swiss Research Foundation NCCR on Neural Plasticity and Repair, the Gottfried and Julia Bangerter-Rhyner Foundation, a Fellowship from the National Science Foundation and the Hans-Eggenberger Foundation. We thank Andreas Brunschweiler and Alessandro Rotta from the ETH Zurich and the occupational therapists and Prof. Dr. V. Dietz of the Balgrist University Hospital, Zurich, for their contributions to this work. Furthermore, we thank Dr. Gery Colombo from Hocoma AG, Volketswil, Switzerland, for his contributions.
Notes
114 HFUC gearbox, Harmonic Drive Inc., Japan.
2Maxon Inc., Switzerland.
a Worst case exoskeleton position.
b Measured with healthy subject.
c Measured without subject (exoskeleton only).
d Stiffness measured at the endpoint by applying 20 N, while the motors are position controlled.
e According to Equation (24), the overall friction torque is τf j ( j ) = sgn ( j )τs j + c j j (1 ≤ j ≤ 4).
f An additional force sensor (6 DOF, JR3 Inc., USA) has been used to measure the breakaway torques. Details can be found in Nef et al. (2008b).
g r cg1 varies with the adjustable upper arm length l u and with the distance d. It is r cg1 = 1/m(r 1a m 1a+ r 1b m 1b) = 1/5.77 kg(0.105 m× 2.28 kg) + (l u− d − q 7 − 0.01 m)3.49 kg). r cg4 varies with the lower arm length l l, but as the handle is lightweight, the variations are very small and therefore neglected.
h The motor current of axis 3 is limited to 3 A. The current for the other axis is limited to 10 A.