Research

Some of what we build is meant to last longer than one season: a benchmark other groups can measure against, software released for other teams to use, and the research that comes out of running a competition system.

The list below collects those contributions. For the papers themselves, see publications .

• Benchmark

  WorkBenchMark

  A LEGO Duplo-based robotic assembly benchmark

  A benchmark for robotic assembly motivated by the RoboCup Smart Manufacturing League, pairing low-level manipulation with symbolic reasoning. It contains 400 tasks across four difficulty tiers, simulation tooling and an Assembly-by-Disassembly baseline. Developed with the MASCOR Institute (FH Aachen) and the Chair of Machine Learning and Reasoning (RWTH Aachen).

  Oral presentation, RoboCup Symposium 2026, Incheon

  • workbenchmark.github.io
  • arXiv:2606.19358

• Software

  ROS 2 CLIPS Executive (ROS CX)

  Rule-based task execution for ROS 2

  Our CLIPS-based executive, packaged for ROS 2 and released for anyone to use. It runs the high-level decision making on our robots and is the piece of our stack most likely to be useful to other groups.

  Viehmann et al., ICAART 2026

  • clips_executive
  • arXiv:2512.12722

• Software

  ROS 2 stack for the Smart Manufacturing League

  Perception, navigation, manipulation and reasoning

  Our current stack is developed in the open: a CLIPS-based executive and goal-reasoning agents, behaviour trees, a fused multi-laser navigation stack, vision and gripper control. Reusable well beyond the competition.

  • github.com/carologistics
  • Software overview

• Software

  Simulation for the league

  Repeatable experiments without physical robots

  A full simulation of the competition environment - first in Gazebo alongside the Fawkes stack, now as a Webots world for ROS 2 - which makes league research reproducible and lets teams develop without hardware.

  • rcll_simulation_webots

• Publications

  Planning and execution research

  Goal reasoning, execution monitoring, multi-robot coordination

  The competition is a testbed for research at both institutes: goal reasoning under uncertainty, execution monitoring, temporal planning and coordinating a fleet of robots on a shared production task. Results feed back into the competition system each season.

  RoboCup Symposium, ICAPS, KI and others

• Teaching

  Student theses

  Bachelor's and master's work on the competition system

  The competition system is a standing supply of thesis topics at both institutes, covering navigation, perception, manipulation and high-level reasoning. Results from that work go back into the robots we compete with.

Working with us

If you would like to use any of this in your own work, please get in touch . Everything we release is open source.