Control of Robotic Manipulator Using ROS

Abstract

This thesis addresses the modeling and control of the Universal Robots UR5e 6-DOF collaborative manipulator. Complete forward and inverse kinematic models are derived using the modified Denavit–Hartenberg convention, alongside a full dynamic model obtained via the Euler–Lagrange formulation. Three joint-space control strategies—pure PD, PD with gravity compensation (PD+G), and Computed Torque Control (CTC)—are designed and comparatively evaluated. The CTC approach leverages the complete dynamic model to cancel nonlinearities and decouple the joints, transforming the system into independent linear double integrators. All controllers are implemented and validated in a ROS/Gazebo simulation environment. Results demonstrate that CTC achieves superior trajectory tracking accuracy and reduced steady-state error compared to both PD and PD+G, confirming the effectiveness of model-based control for the UR5e manipulator.