CargoBOT: Design, Test and Validation of a smart machine for automatic container unloading

Seventy years after its invention the container still is the most used transportation unit in the world, in fact the number of containers moved every year is increasing more and more. Yet the innovation around the container industry remain very little compared to the magnitude of its market. Many tasks are even the most tiring and dangerous are performed manually. In particular, one of the biggest question mark of many logistics player is how to efficiently unload non-palletized containers. In this thesis, an innovative way to automatically unload container is presented. Starting from an analysis of the available solution on the market: a new prototypal design is proposed. A prototypal robotic cell is firstly developed and tested in a simulated environment, then upon reaching good performances is deployed in the real world. This cell uses a sensing unity mounted on the robot leverages the huge potentialities of 3D vision and AI to recognize objects. In fact, is able to autonomously recognize bags from a random layout, estimate their pick point, create the robot trajectories accordingly and palletized the picked object in an ordered way.