Position control of a manipulator robotic arm considering flexible joints driven by a dc motor and a controlled torque by a mr-brake

Abstract

This paper presents a two-degree-of-freedom robotic arm design with flexible joints driven by a DC Motor and controlled by a Magnetorheological (MR) Brake, considering a feedback control. The MR Brake is used to provide adjustable constraints in motion of the manipulator and compensate overshoot by interactions between the robot's links and flexible joints of the motor drive mechanism. The torque of the MR Brake is obtained by the Radial Basis Function Neural Networks (RBFNN), which is a widely used class of neural networks for prediction or approximation of function. The RBFNN provides the nonlinear curve of hysteresis of MR brake to use torque. Two controllers were proposed to control the manipulator. The first one is obtained by feedback linearization control with the objective to remove the non-dependent terms of the state space equation. The second one is the feedback control obtained using the State-Dependent Riccati Equation (SDRE) with the objective of controlling the position of the manipulator and the torque applied on the MR brake. The numerical simulation results showed that the proposed control using both signal feedback linearization control and a feedback control signal by a DC Motor and MR Brake is effective to control the flexible joint manipulators.