GRASS: Grasping into the Shelf

Project Description

The GRASS project aims to automate the order picking process in a warehouse using a mobile picking robot. The focus is on the intelligent grasping of various product types from the shelf and their robust placement on a target pallet.

CERI's research has focused on two areas:

  1. Developing a pipeline for the synthetic generation of image data enriched with real-world textures.
  2. Object detection through a combination of machine learning methods and the geometric properties of the goods.

Short Facts

  • Duration: Feb '21 - Jun '25
  • Funding: StMWi Bayern
  • Partner: Linde Material Handling

Project Team

Publications

 Synthetically Generated Image Data

 

 

Automatically Generated Segmentation

 

 

Grasp Poses via Depth Images