Performance verification of different control schemes in human lower extremity rehabilitation robot

Abstract

Many people experience temporary or permanent impairment sometime in their lifetime. Exoskeleton robot-based exercise machines for neurorehabilitation offer an attractive alternative for many disabled people. The performance of exoskeleton robot-based physiotherapy depends on the effectiveness of the robot control scheme. This paper examines the performance of three commonly used control algorithms in serial manipulators. Before examining the performance of the control schemes, it is essential to establish a model for plant dynamics. A dynamic model of the human lower extremity (HLE) is developed first. The Lagrange energy method is used for developing the HLE model. Since the dynamics of a serial manipulator is inherently nonlinear, a nonlinear control scheme would appear effective in dealing with the nonlinear robotic system. Two nonlinear control schemes (computed torque controller and sliding mode controller) and one linear controller (PID) are used to control the robot. The performance of the three control schemes is examined when disturbance torques and/or variations in robot parameters are present. The pros and cons of linear and nonlinear control schemes are discussed. Finally, a detailed discussion on controller selection criteria is presented, including how it can facilitate the selection of a suitable controller.