ICRA 2023 Virtual Manipulation Challenge: Bin Picking, Assembly, and Stacking

The Virtual Manipulation Challenge aims to reduce the gap between robotic simulation and real-world setups, providing the possibility to train robot behavior for complex manipulation scenarios first in simulation and later transferring the learned policy seamlessly to the real-world setup. The challenge addresses three different manipulation tasks frequently occurring in automation industry: (i) picking individual objects from heavily cluttered bins, (ii) assembly by screwing and plugging, and (iii) stacking. These scenarios offer the opportunity for researchers and industrial professionals in the field of autonomous robotic manipulation to apply their approaches to ambitious real-world industrial use cases provided by internationally renowned industrial sponsors.

The challenge and award ceremony will be held at the 2023 IEEE International Conference on Robotics and Automation (ICRA 2023) on May 29 - 03 June, 2023 in London, UK.

1. Bin picking: This track involves tasks on machine tending and kitting (compiling all parts required for a specific assembly step), which are modeled after use cases from our project partners Käpple Qualitätsleister GmbH and Daimler Truck AG as well as the external sponsor Magna.
   
   Learn more about the Bin Picking Track
   
   
2. Assembly: This track consists of two force and/or vision guided assembly tasks by Daimler Truck AG that involve tightening a screw and plugging fuses.
   
   Learn more about the Assembly Track
   
   
3. Stacking: This track is concerned with the tasks of parcel loading and ad-hoc item packing, which represent two current challenges in logistics inspired by Vanderlande Industries B.V.
   
   Learn more about the Stacking Track

In each track participants are asked to develop a robot controller which can solve the respective task with maximum performance and robustness.

For each track we will provide a parameterizable Docker container running the respective simulation setup. This Docker container will be publicized by 31^th January 2023 for the contestants to set up, train and validate their robot controllers for the given tasks. Participants should prepare a compiled Docker image of their robot controller, which interacts with the provided simulation container via ROS. The contributed robot controller will be evaluated before the ICRA 2023 conference by the organizers against novel test cases run in the simulation container.

Who can participate?

This competition is open to everyone working or willing to work on the respective tasks of the individual competition tracks. Researchers from industry and academia working on these tasks are especially encouraged to participate in the challenge. We also welcome commercial solutions to the tasks at hand in the competition.

How to participate?

1. Register here.
2. Choose at least one of the three given tracks and download its Docker environments (download links will be sent to you via email by 31.01.2023).
3. Choose or develop a method and prepare your routines of the respective task in a Docker container runtime environment (a template container will be shared with you as well).
4. Submit your Docker container according to the instructions shared with you by email. Each registered team will receive a dedicated individual upload link.

10^th January 2023: Registration for the challenges opens.

31^th January 2023: Docker containers for the different tracks will be available for download.

14th April 2023: Registration for the challenges closes.

7^th May 2023: Deadline of solution submission.

30^th May 2023 - 2^nd of June 2023: All teams gather at the official competition days at ICRA 2023, discussion of results and announcement of winners, award ceremony (concrete date for each track will be provided soon)

The Virtual Manipulation Challenge is organized by the research project Sim4Dexterity. The Sim4Dexterity project has received funding from the German Federal Ministry of Education and Research (BMBF) through grant 01IS21061 (Sim4Dexterity).