Robotic devices have been widely used in gait rehabilitation to help patients improve or regain the ability to walk. While rehabilitation robots can deliver high-dosage and high-intensity training, their efficacy is often limited due to their friction, bulkiness, and inertia. To make such devices behave smoothly and gently to the human motion, several hardware and control solutions have been proposed to improve their mechanical compliance. In fact, a robotic interface able to render low impedance can enhance active user participation an achieve low robotic interference. 

For instance, most rehabilitation robots rely on the intrinsic backdrivability of the actuation units, which can be improved through model-based mass, friction and inertia compensation models. In other works, robot transparency is obtained through appropriate force/torque sensing to minimize the interaction forces at the interface points with the human leg. 

In this project, you will first review lower-limb rehabilitation robots and available strategies to improve their mechanical transparency. Then, you will employ a highly customized treadmill-based gait rehabilitation robot (Lokomat, Hocoma AG, Switzerland) and conduct its system identification through a battery of testbed experiments. With these results you will be able to develop (and possibly validate) some feedforward-compensation algorithms to compensate for the robot mass, friction and inertia.