Development of a multi-environment platform composed by a robot and an autonomous guided vehicle

The thesis work was developed in collaboration with Comau S.p.A. Comau is a company that is part of the FCA Group, specialized in industrial automation and in particular on welding robots. The aim of the thesis was to create an autonomous system, close to the concept of industry 4.0, that is able to manage and coordinate the movements of two robots which interacts with each others in the same environment. In particular, the developed system manages the communication between an anthropomorphic robot and an autonomous guided vehicle (AGV) in order to perform the coordinated movement of both the robots and execute a predefined task. The anthropomorphic robot considered in the system is the robot e.DO, a modular educational robot developed by Comau S.p.A. The robot is a manipulator with a spherical wrist and it is composed by six joints, the first three belonging to the arm and the lasts to the wrist. The AGV was not available during the development of the thesis, for this reason, it has been performed a gazebo simulation of the system with a differential drive mobile robot in the role of a collaborative assistant for e.DO. The task consists of releasing an object grasped by e.DO inside one of the two boxes positioned on the base of the AGV and to coordinate the movement between the two robots, a camera and a laser sensor have been used. The camera aims at correcting the trajectory of the AGV and at identifying the point in the space where to release the object inside the correct box, while the laser scan sensor has been used in the gazebo simulation to measure the AGV distance from e.DO and to check if the AGV is positioned correctly in the target release zone. In particular both the configurations for the camera have been analyzed: camera-to-hand and camera-in-hand, and the system has been developed in order to works with both the positions of the camera. To coordinate the movement of the gripper above the target box, image processing operations are performed starting from the image provided by the camera. Depending on the type of the configuration and on the image processing techniques, different strategies have been implemented for the vision system, which are based on the information extracted from the image and on the use of the Comau library that allows to move e.DO. The whole system is managed by an external raspberry pi controller which communicates with the robot and with the autonomous guided vehicle. In order to do that, it has been developed a specific ROS package, which runs on the raspberry pi controller, that manages the two robots and synchronizes all the movements to complete the task.