The ARM Challenge is a global competition that tasks university teams with solving complex manipulation problems using real robotic arms. Participants must bridge the gap between simulation and reality, designing algorithms and simulating them in MathWorks MATLAB software, and then adapting to a real- environment with a real objects and a Universal Robot UR7e Robot equipped with a camera. Students are challenged to address constraints like sensor noise, reach limitations, and time pressure.
From Simulation to Reality
For the AGH team, prior experience with MATLAB and Universal Robots gave them a strong foundation. But as they quickly discovered, real-world robotics is a different game. “In simulation, everything works perfectly,” they noted. “But in real life, we had to learn how to properly manipulate objects, deal with robot singularities, and adjust for noisy camera data—especially from the depth sensor.”
Strategy and Execution
Their strategy was both methodical and adaptive:
- Stick to the rules: They ensured every move aligned with competition guidelines.
- Leverage predictability: For static objects, they used known positions to their advantage.
- Adapt for dynamic objects: They developed a consistent approach—moving to a fixed position above the object, using camera data to estimate orientation, and adjusting the pickup accordingly.
- Bold final move: In their last run, they implemented a new strategy—picking up a small basket containing multiple objects of the same class and dropping it into the correct bin. It was a risk, but it paid off.
Learning Under Pressure
Though they had experience with UR3 robots and MATLAB at their university, this was their first time applying those skills in a real-world competition setting. With only four days and limited access to the robot, they had to work fast, think critically, and adapt on the fly.
Interested in competing the next Robocup ARM Challenge? Registration opens in November here.