POSE ESTIMATION AND CONSTRAINED CONTROL OF A ROBOTIC MANIPULATOR IN OPERATIONAL SPACE

This work lies within the field of industrial robotics, specifically focusing on motion planning and control of articulated manipulators using open-source tools based on ROS2. The project concentrates on operational space control of a fixed-base manipulator by integrating a trajectory planning tool for trajectory planning with a visual localization system based on SLAM to estimate the end-effector pose. The goal of this thesis is to develop an operational space controller that uses the error between the desired end-effector pose and the one estimated through SLAM to generate torque commands for trajectory tracking. This work includes the integration of an RGB-D camera mounted on the robot, the Niryo Ned2, the implementation and configuration of a SLAM system for pose estimation, trajectory planning with the MoveIt2 tool, the design of a custom controller as a ROS2 plugin, and the execution of simulations to evaluate the trajectory tracking performance. The results confirm the feasibility of the proposed approach and highlight both its limitations and its potential for future developments in real-time visual control scenarios.