Mobile-Menu

Put into a trot Four-legged robot learns to walk energy-efficiently

Source: TU Munich | Translated by AI 2 min Reading Time

Related Vendors

TU München
FAULHABER Drive Systems
Taiwan Machine Tool & Accessory Builders' Association (TMBA)
Wuxi InfiMotion Technology Co., Ltd.

When four-legged creatures accelerate their pace, they naturally start trotting because it saves energy. A robot from TU Munich can now do this too.

This is the four-legged robot dog Bert. With it, researchers at TU Munich have now achieved a breakthrough in understanding the most energy-saving movement methods ...(Image: TU Munich / A. Heddergott)
This is the four-legged robot dog Bert. With it, researchers at TU Munich have now achieved a breakthrough in understanding the most energy-saving movement methods ...
(Image: TU Munich / A. Heddergott)

That quadrupeds switch to the energy-saving trot at a certain running speed has been known for over 40 years, say researchers from the Technical University of Munich (TUM). Now, with a view to sensor-based robotic systems and "intelligent" assistive systems, it has been possible to transfer this behavior to the movement of robots. Experts speak of intrinsic dynamics, which humans and animals use for their energy-efficient movements. For example, they adjust the stiffness of their muscles when running on harder surfaces. These instinctive adjustments occur automatically and are difficult to identify, it is said. This applies to both humans and complex robotic systems.

On the trail of the most economical movement with Bert

However, with a new tool from TUM, this knowledge gap has now been closed. This makes it possible to determine which movements of a system are especially economical. An important test subject is Bert. He is a four-legged robot that looks like a small dog. Bert was designed at the German Aerospace Center (DLR). This research focused on "efficient and versatile legged locomotion" and is funded by an ERC Advanced Grant from the EU. The involved experts identified six motion patterns for the mechanical quadruped that require no energy in a frictionless world, as emphasized. These partially correspond to known gaits of quadrupeds (walking, trotting, or hopping). This has confirmed the hypothesis that efficient gaits could be achieved by exploiting natural oscillation patterns.

Robot race confirms the hypothesis in practice

To implement these movements in a normal system (i.e., with friction), a computer-controlled controller was implemented to give an impulse at the right moment, it is said. You can imagine it similarly to a child sitting on a swing who receives an energy impulse from their parents at the highest point to push them. But there's a difference! The human doesn't require a lot of equations in their head for exact timing—it's done intuitively. However, the robot needs to be taught the correct rhythm. In a race between three Bert models, the correctness of the hypothesis was revealed. The robotic dog that mastered the method of intrinsic movement moved more dynamically and rapidly, almost jumping, compared to its siblings, which were programmed with more conventional movement patterns.

Subscribe to the newsletter now

Don't Miss out on Our Best Content

By clicking on „Subscribe to Newsletter“ I agree to the processing and use of my data according to the consent form (please expand for details) and accept the Terms of Use. For more information, please see our Privacy Policy. The consent declaration relates, among other things, to the sending of editorial newsletters by email and to data matching for marketing purposes with selected advertising partners (e.g., LinkedIn, Google, Meta)

Unfold for details of your consent

Related content