Mari4_Yard | Mobile Manipulator for intra-logistic operations in shipyards

The transportation of raw materials and manufactured parts in shipyards are still heavily reliant on human operators. This transportation is typically performed by hand or using self-propelled, pulled, or pushed platforms. However, since these logistic tasks are dull, dirty, and dangerous for the human operator, and due to the ageing of the European population, it is important to empower the current human workforce to perform other tasks. Therefore, there is a high interest in the shipbuilding sector to automate its intra-logistic operations.

To answer these challenges the Mari4_Yard Project (https://www.mari4yard.eu/) proposed the development of a Mobile Manipulator that allows picking individual parts from containers. The biggest advantage when compared with traditional AGVs and AMRs is that Mobile Manipulators combine the capacity to transport the load on top of the mobile platform with the manipulation dexterity of industrial robotic arms.

For localization, the robotic system combines odometry data with the point clouds from two Sick Lidar lasers, allowing the extraction of the environment’s natural contour, as well as the detection of artificial markers. These same lasers are also used to ensure the safety of the operation by detecting the presence of obstacles and triggering the robotic platform to reduce its speed.

A Task Manager acts as the local orchestration module, being responsible for the supervision of the execution of different modular robotic skills, such as Drive Skill and Pick and Place Skill. The Task Manager is also responsible for capturing operational data from the robotic system, to be integrated and processed in the upper layers of the manufacturing stack.

The robot system also allows the use of an Augmented Reality system (AR) based on Microsoft HoloLens to improve human-robot interaction by implementing a codeless production task programming approach. During these interactions, the AR system also monitors a 3D volume (safety zone) around the manipulator. Whenever safety zones are breached, the system triggers the robot to change its speed (allowing the object to move away from the robot’s path) or even to come to a complete stop, thus preventing imminent collisions.

A Perception Module implemented in the Mobile Manipulator enables the robotic arm to pick a wide set of parts stored inside a box or a bin. It is responsible for accurately perceiving and segmenting objects in cluttered picking scenarios, estimating the position of the object to grasp based on a reference point cloud, computing grasp pose candidates, and choosing the best grasp pose based on effort, joint space limitations, and potential collision with objects.

The Mari4_Yard Project has received funding from the European Union’s Horizon 2020 research and innovation programme under Grant Agreement n° 101006798.