Manipulability based model predictive control of rehabilitation robot

One challenge with rehabilitation exoskeletons is the potential for reaching singular configurations, reducing efficiency. Additionally, paths generated for an exoskeleton may not always exhibit optimal manipulability and dexterity, unlike healthy humans who perform tasks with maximum manipulability. This paper considers a multi-degree-of-freedom model for the exoskeleton robot, deriving its kinematic and dynamic equations. The robot's kinematic manipulability is formulated based on the Jacobian, and Model Predictive Control (MPC) is employed for control. The novelty lies in incorporating the cost function related to the robot's kinematic manipulability alongside other cost functions within the MPC framework. Dynamic simulations evaluate this approach, showing that the manipulability criterion conflicts with tracking error. This research demonstrates that using the manipulability index as a constraint or part of the cost function in MPC can help prevent the robot from reaching singular points and enhance manipulability and dexterity in hand rehabilitation.

کلیدواژه ها:

Rehabilitation robot ، model predictive control ، manipulability