Constraints – three different usecases
CoboSort slution addresses the picking of garments that are collected inside logistic trolleys in three different conditions: items packaged, partially packaged (broken envelopes), and bulk. Inside the trolleys, the garments are systematically collected to use all the volume available, but a random sequence is used to fill the trolley.
Smart vision system with AI for garment recognition
Adoption of segmentation techniques and data labeling, to traing the machine learning models. The expected results are the singularization and location of the items. As a result, the grasping point is returned to the cobot. An approach to integrate continuous improvement of the machine learning model by human intervention is also provided.
Gripper system for garment grasping
Development of a dedicated gripper system for gentle and soft grasping of the three different use cases selected. Existing gripper modules are enanched to ensure an effective grasping action: repetivie and efficient.
Human-robot collaborative enviroment
Human-robot collaborative architecture for easy and safe integration in human enviroments. Modular configuration for vision and gripper systems, as weel as for safety layer for easy customization of the solution.
