A ROS-based Simulation Framework for Industrial Robotics Pick-and-Place Application: Integration of Vision-based Pose Estimation and 3-DOF Inverse Kinematics

The efficiency of industrial automation depends heavily on the precision and reliability of robotic pick-and-place applications. This paper presents a complete simulation framework designed to develop and validate such applications in the ROS (Robot Operating System) environment. We perform a closed-form inverse kinematics analysis for the 3-degree-of-freedom OpenManipulator-X, providing the critical transformation from task-oriented Cartesian space to the manipulator's joint space. We apply this kinematic solution to the robot using an OpenCV-based image processing algorithm, enabling the robot to dynamically target a planar object position estimation. Within the ROS architecture, pose data is broadcast from our vision node and subscribed to by the inverse kinematics node, which calculates the required joint configurations. To achieve the smooth and precise motion required in industrial environments, a cubic trajectory is generated for joint space, and a PID controller is implemented to ensure the end-effector accurately grasps the target object. All simulation studies are conducted in the Gazebo environment, demonstrating a robust and integrated process from visual detection to successful manipulation.