The rise of robots, particularly Boston Dynamics' Spot, has highlighted the remarkable capabilities of quadrupedal systems. These advanced robots excel at climbing stairs, navigating small obstacles, and moving across uneven surfaces. However, ladders remain a significant challenge, especially in industrial settings where they are commonly found.
ETH Zurich, a leader in innovative quadrupedal robot research, has unveiled a promising solution. The university points out that previous approaches to addressing ladder-climbing have primarily focused on bipedal humanoid robots and specialized ladders, which ultimately fell short in speed and efficiency.
The latest development involves the ANYMal robot, a product of the ETH Zurich spinoff ANYbotics. The research team equipped the quadruped with specialized end effectors designed to grasp ladder rungs. The key to this advancement lies in reinforcement learning, enabling the system to adapt to the unique characteristics of various ladders.
“Our research broadens the potential applications for industrial quadrupedal robots, extending beyond standard terrain inspection to include more complex infrastructural features,” the researchers state. “This highlights the interplay between robot design and control strategies when executing intricate tasks.”
The combined system achieved an impressive 90% success rate in navigating ladder angles between 70 and 90 degrees. Furthermore, it demonstrated a remarkable climbing speed increase of 232 times compared to existing state-of-the-art systems. This innovative system can also self-correct in real-time, adjusting its movements if it miscalculates a climb or misjudges the timing of a step.