The Case for Biomimetic Machines

by Robert Katzschmann, Assistant Professor of Robotics & TED Fellow, 14 April 2022
Robert next to the TED logo
Photo credit: Ryan Lash / TED

For my recent TED Talk (April 10), I introduced the audience to new approaches in robotics engineering.

Modern engineering relies on rotating motors, noisy rotating motors that not only power our machines, but also produce noise pollution. The natural world and the man made environment equally suffer from the noise of machines. Whales lose their navigation, humans become stressed and depressed. And that is only the tip of a melting iceberg.

In contrast, nature hums along smoothly, gliding, wriggling, pulsating through its processes sans the rattling of rigid parts, loudness of surfaces rubbed together. We can and should take cues from nature on how to make machines: we can learn to minimise friction, self-lubricate, and last for billions of movements.
Robert on a stage with the picture of two whales behin him
Photo credit: Ryan Lash / TED
What if we rehauled our approach and ditched rigid synthetic materials in favour of biomimetic machines? These would be machines devised from soft organic materials—adaptive and safe for a variety of quotidian tasks.

Imagine a submarine or a boat that propels like a fish, swishing tails from side to side? To mimic these natural movements, we would replace rotating motors and fast spinning propellers with artificial muscles. At ETH Zurich and MIT our research teams are already on the job. Meet SoFi, a biomimetic robot fish. The first of her kind, fully untethered, she can explore the ocean without the need for propellers. Instead she pumps water back and forth inside her highly deformable tail, whose movements imitate the swimming motions of a real fish. This naturalistic design allows her to closely monitor aquatic life.
There’s a catch: SoFi drains battery quickly—and she is still loud, because of the noisy motor and pump she needs to operate. There’s also a solution: we’re replacing said noisy motorised pump with silent artificial muscles that directly and efficiently transform electrical energy into contractions. This large muscle is created from a stack of flexible plastic sheets that can contract and relax—like a thigh muscle.

These plastic sheets are laminated to create pouches, spray-coated with a conductive ink, and then filled with oil. The conductive ink becomes a flexible electrode that squeezes the pouches when a voltage is applied.

Combining these contracting pouches in opposition to each other, we arrive at the infrastructure for a new robotic fish. A hull and spine is now 3D printed from soft materials, and the two muscles are inserted. The result? SoFi now swims silently by alternating the electrical current of these muscles, one side first, then the other…
Robert on a big stage, before him the audience
Photo credit: Ryan Lash / TED
Robert in discussion with another person
Photo credit: Bret Hartman / TED
Robert on a walk next to an african man
Photo credit: Stacy McChesney / TED
A benefit of software designed performance is the elimination of tedious trial-and-error fabrication. Instead, the software rapidly simulates and optimises before we build the robot. So we are able to smoothly combine the shape and muscle of any predefined design to create a new swimmer. Optimisation explores the continuous design space further: emerging with creatures we have not seen or imagined before.

This all sets a precedent for robots to be made of living muscle cells, and thus not only having the ability to heal themselves, but also proliferate. We are already able to print, grow, and stimulate muscle tissues so they contract. We are now working on scaling them up, experimenting with control, and keeping these muscles alive by feeding them nutrients.

A living muscle outside of a supporting organism (and thus a protective immune system) is a fascinating challenge. Artificial muscles are made from non-degradable rubber or plastic, real muscles on the other hand, derive from natural resources and can efficiently convert chemical energy into mechanical energy. A biohybrid robot, that mimics nature in its operation, could well decrease emissions and pollution.
This science is no longer in the realm of speculation, engineers and researchers already have the sufficient knowledge and are pushing forward. Our team at ETH is already rethinking every machine ever invented from a holistic perspective. Not only are we building noiseless biomimetic machines—we are also rethinking what machines should do and should not do. We want machines that safely integrate into our lives, improve our quality of life, and don’t pollute the world with loud noises. We want to reimagine machines that break with all conventions, take on novel forms we never thought of, and reach places never been.
Selfie of Robert with other people holding a sign that says Fishhhhh
Photo credit: Robert Katzschmann
Portrait of Robert Katzschmann
Photo credit: Ryan Lash / TED

About the author

Robert Katzschmann is an Assistant Professor of Robotics at ETH Zurich. Robert earned his Diplom-​​Ingenieur in 2013 from the Karlsruhe Institute of Technology (KIT) and his Ph.D. in Mechanical Engineering in 2018 from the Computer Science and Artificial Intelligence Lab (CSAIL) at the Massachusetts Institute of Technology (MIT) in 2018. Robert worked on robotic manipulation technologies as Applied Scientist at Amazon Robotics and as CTO at Dexai Robotics. In July 2020, Robert founded the Soft Robotics Lab at ETH Zurich to push robots' abilities for real-​​life applications by being more compliant and better adapt to their environment to solve challenging tasks.