Researchers at the Technical University of Munich (TUM) have developed a new approach to improve energy-efficient robotic movements by emulating the natural rhythmic patterns observed in humans and animals. Under the leadership of Prof. Alin Albu-Schäffer, the team has succeeded in mathematically calculating “intrinsic dynamics,” enabling robots to move with greater speed and fluidity. This breakthrough stems from efforts funded by the European Research Council through an ERC Advanced Grant.
Intrinsic dynamics describe the natural energy-efficient movement patterns seen in biological systems, such as humans adjusting muscle stiffness based on surface conditions or animals adopting energy-saving gaits. For example, quadrupeds naturally transition from walking to trotting as their speed increases, a behavior observed to minimize energy expenditure. Prof. Albu-Schäffer and his team have translated these principles to robotic systems, focusing on efficient and versatile legged locomotion.
At the core of the research is BERT, a four-legged robot resembling a small dog, designed by Prof. Albu-Schäffer at the German Aerospace Centre (DLR). Using a newly developed computational tool, the researchers identified six energy-efficient movement patterns for BERT, some of which correspond to natural gaits such as walking, trotting, and hopping. These movements are modeled to exploit the robot’s natural oscillation patterns, confirming hypotheses about the efficiency of biologically inspired gaits.
The team implemented a computer-controlled regulator to integrate these movements into practical robotics. This regulator ensures precise timing of impulses to maintain rhythmic motion, akin to a parent pushing a child on a swing at the right moment. Doctoral student Annika Schmidt noted that while humans instinctively time such actions, teaching robots this capability required years of research to establish computational methods.
The success of the approach was demonstrated through tests involving multiple BERT models. Robots programmed with intrinsic movement patterns showed significantly faster and more dynamic locomotion compared to counterparts using traditional movement algorithms. This development marks a step forward in creating robots capable of natural and efficient movement, paving the way for broader applications in robotics.
