Individualized production places increased demands on logistics. In particular in car manufacturing, each car on an assembly line is different and the automotive logistics must ensure that the right parts are delivered for assembly at the right place in the right moment.
Part sequencing, the collection of the right variants of one category of parts in the order of car assembly, and kitting, the collection of parts of different categories that are assembled into a particular car, are the logistic processes to developed to address these demands. These operations are performed in the vicinity of the assembly lines by logistic workers who receive a “shopping list” of parts and collect the required parts in a “part supermarket”. The manual part collection is cost intensive and error-prone.
The goal of the proposed project is to realize a mobile robot that performs sequencing and kitting tasks for automotive logistics in a workspace shared with human co-workers. Parts of the kit will be assembled by the robot and the remaining parts will be collected by the human logistic workers. This will allow both to bring in their strength.
For part manipulation, it is necessary to detect known automotive parts in storage containers and transport boxes, and to estimate their pose. Challenges include varying lighting conditions and surface properties of typical automotive parts, such as shiny metal surfaces, black surfaces, or transparent materials. Furthermore, other parts and packaging material may occlude parts partially.
Parts must be grasped firmly to pick them from the storage containers. Depending on the properties of the parts, such as size, shape, weight, and surface material, different gripping technologies might be necessary. Space restrictions from the packaging and neighbouring parts may severely limit access to the part. Despite these challenges, parts shall be picked reliably.
The picked parts must be placed at defined slots in transport boxes, in defined poses. Re-grasping might be necessary to bring the part from the picking pose into the placing pose. Placing must be done gently to avoid part damage.
The robot will pick parts in a “supermarket” along with human co-workers. It must acquire a map of this environment, continuously localize with respect to this map, and reliably perceive dynamic objects, such as co-workers, carts, other robots, obstacles on the floor, etc.
The robot must navigate to the storage containers from which the parts are picked and to an exchange station for delivering filled kits in exchange for empty kitting boxes. It must avoid all static and dynamic obstacles and will adapt its route according to the situation.
The developed sequencing and kitting solution will cope with uncertainty and environment dynamics. It will be easy to adapt the system to new parts by non-expert users. The mobile robot will relieve the human worked from repetitive and physically demanding tasks, save costs and increase reliability of part collection. This will strengthen the competitive position of the European car manufacturing industry. The developed solutions will be general enough to be applicable to many related logistic operations.