Despite better individual components, robots lag behind animals in terms of overall performance, indicating that in the future the field of robotics will focus on better system integration and control.

Robotics engineers have been working for decades and have invested millions of dollars to create a robot that can walk or run as much as an animal. However, the fact remains that many animals have skills that are impossible for today’s robots.

“The wildebeest can migrate thousands of kilometers over rugged terrain, the mountain goat can literally climb a rock, finding handholds that don’t seem to exist, and the cockroaches can lose a leg and not slow down,” says the doctor. Max Donelan is Professor in the Department of Biomedical Physiology and Kinesiology at Simon Fraser University. “We don’t have robots with this much endurance, dexterity and strength.”

Comparing robots to animals

An interdisciplinary team of scientists and engineers from leading research universities has completed a detailed study of various aspects of robot operation and compared them to their animal counterparts, with the aim of understanding and quantifying why robots are lagging behind animals. Science Robotics. . The paper found that biological components performed surprisingly poorly compared to manufactured parts, based on measurements used by engineers. However, animals excel in the integration and control of these components.

In addition to Donelan, the team includes Dr. Sam Bearden, Associate Professor in the Department of Electrical and Computer Engineering at the University of Washington; Tom Libby, Senior Research Engineer, SRI International; Kaushik Jayaram, Paul M. Rudy Associate Professor of Mechanical Engineering at the University of Colorado Boulder; and Simon Sponberg, assistant professor of physical and biological sciences at the Georgia Institute of Technology.

The researchers examined one of five different “subsystems” (Power, Frame, Actuation, Sensing and Control) that come together to create a working robot and compared them to their biological counterparts. It was previously believed that the superiority of animals over robots must be due to the superiority of biological components.

“With few exceptions, engineered subsystems perform as well as, and sometimes radically outperform, biological equivalents,” says Libby. “But when you compare animals to robots across the system in terms of movement, it’s very, very clear that animals are amazing. And robots still have to catch up.”

Ideas about the development of robotics

More optimistically for the field of robotics, the researchers noted that if you compare the relatively short time it took robotics to develop its technology with the countless generations of animals evolving over millions of years, progress is actually quite rapid. .

“It will progress faster because it is not evolution-oriented,” says Bearden. “Although we can pretty much change the way we design robots and learn something from one robot and transfer it down to other robots, biology doesn’t have that ability. So there are ways we can go much faster than evolution in building robots, but evolution has a huge advantage.”

More than an engineering challenge, efficient treadmills offer countless potential applications. This technology has many potential applications for solving last-mile delivery challenges in a human-made world that is often difficult for wheeled robots to navigate, search in hazardous environments, or handle hazardous materials.

Future directions in robotics

The researchers hope that this research will guide the future development of robotics by focusing not on building better hardware, but on understanding how to better integrate and control existing hardware. Donelan concludes: “As technology learns the principles of integration from biology, running robots will become as efficient, flexible and robust as their biological counterparts.”