Why do animals outrun robots?

Scientists can engineer machines that meet or exceed the performance of their biological counterparts in many areas, however animals still outperform robots in locomotion—running, flying swimming, walking and more.

Simon Fraser University (SFU) biomedical physiology and kinesiology professor and chair Max Donelan has a particular research interest in learning why animals outrun robots. He leads the SFU Locomotion Lab which studies how people and other animals move—and applies that knowledge to help society. He also co-leads the SFU WearTech Labs, a Core Facility that researches and develops wearable technology to improve lives.

Donelan and his research team have invented exoskeletons that harvest electrical energy from movements, devices that stabilize people as they walk, and an iPhone app that controls people’s running pace with music. He has published widely about the ways humans, animals and robots move and has been featured in national and international media.  

In a recent study published in Science Robotics, Donelan and colleagues observed how animals and robots performed in important dimensions of agility, range, and robustness and in the in the five subsystems critical for running: power, frame, actuation, sensing and control.

The paper, Why animals can outrun robots, discusses biology’s advantage over engineering, and identifies four fundamental obstacles that roboticists must overcome.

The research team included Sam Burden, associate professor in the Department of Electrical & Computer Engineering at the University of Washington; Tom Libby, senior research engineer, SRI International; Kaushik Jayaram, assistant professor in the Paul M Rady Department of Mechanical Engineering at the University of Colorado Boulder; and Simon Sponberg, Dunn Family associate professor of Physics and Biological Sciences at the Georgia Institute of Technology.

We spoke to professor Donelan about his research.

Your article talks about biology’s advantage over engineering when it comes to movement. How do animals and machines compare in other ways?  

Robots are becoming amazing at moving about. However, animals still drastically outperform robots along pretty much every performance dimension we can consider. We used agility, range and robustness, but you might imagine others as well. This performance gap is seen in events like robot soccer where the best robot soccer players can be easily bested by human toddlers that are just learning to play the game. This is in sharp contrast to more intellectual human activities, like chess, where the best human chess players to have ever lived, and who have trained their whole life, are now easily bested by computer programs.

What did you conclude about animals versus robotics? What are the obstacles that roboticists need to overcome?

We wondered if the difference in overall performance might be explained by biological runners being built out of superior components like stronger muscles or faster nerves. Afterall, this biological superiority is often the story that biology teachers tell their students. However, when we studied this systematically, we found that for five subsystems critical for running—power, frame, actuation, sensing and control—the performance of engineered components matched or exceeded their biological counterparts. So, roboticists do not need to put their energy into discovering or building better components. Instead, they need to get better at integrating the hardware in these subsystems into a higher-performing whole, and learning how to better control the resulting system. 

Why does the research team study other animals—not just humans—to design efficient robots? How does learning how a kangaroo jumps or a how mountain goat climbs apply to robotics?

Roboticists use many different types of animals as bioinspiration for building robots. Just amongst my group of co-authors, we have studied cockroaches for their robustness, geckos for their ability to stick to walls, flies to understand how lift can be generated from wings that move, and kangaroos for how they use their tail as a fifth leg. They key is to learn principles from biology that can be applied to make robots better at their job. For example, a robot built from the principles underlying cockroach running might be better suited than a humanoid robot for search and rescue tasks in the rubble after an earthquake. Interestingly, we do not just make robots better by understanding biology, we also understand biology better by building robots.

In your work with the Locomotion Lab and the WearTech Labs, you develop, test and commercialize wearable technologies designed to improve quality of life. Can you tell us more about the work of WearTech Labs?  

WearTech Labs is just getting off the ground. It is a $20 million state-of-the-art facility staffed by scientists and engineers experienced in making and testing wearable technology. This is wearables broadly defined—things like smart watches—but also includes running shoes, technical apparel, and exoskeletons. We have a 3D printing lab that can print all kinds of materials (including metal), a smart textile lab that can knit electronics into clothing, a motion capture lab that can measure the energy used by individuals as they wear new inventions and an environmental lab that can simulate the temperature, humidity and wind conditions of pretty much anywhere on earth. Our mission is to work with both industry and academic labs to help develop and test their ideas, and together make technology that improve lives.

How did you become interested in this area of research and what makes you most proud about this work?

Thanks to my father, I had some natural skills at math and physics. However, my initial passion was athletics. As an undergrad student, this led to me combining these interests by doing research in a biomechanics lab. Since then, I have continued to balance both an interest in how animals work, with how to engineer and build things that either work like animals, or help animals work. As for what I am most proud about for this project, it is how my co-authors and I collaborated in bringing our individual strengths to bear on an interesting problem, made some novel insights and had a ton of fun doing it.

For more: Read the SFU News story, and listen to Max Donelan’s interview on CBC’s Quirks and Quarks.