The robotic leg jumps across different terrains. (Image: Thomas Buchner / ETH Zurich und Toshihiko Fukushima / Max-Planck-Institut für Intelligente Systeme)

Inventors and researchers have been developing robots for almost 70 years. To date, all the machines they have built — whether for factories or elsewhere — have had one thing in common: they are powered by motors, a technology that is already 200 years old. Even walking robots feature arms and legs that are powered by motors, not by muscles as in humans and animals. This in part suggests why they lack the mobility and adaptability of living creatures.

A new muscle-powered robotic leg is not only more energy efficient than a conventional one, but it can also perform high jumps and fast movements as well as detect and react to obstacles — all without the need for complex sensors. The new leg has been developed by researchers at ETH Zurich and the Max Planck Institute for Intelligent Systems (MPI-IS) in a research partnership called Max Planck ETH Center for Learning Systems, known as CLS. The CLS team was led by Robert Katzschmann, Soft Robotics Lab at ETH Zurich and Christoph Keplinger at MPI-IS. Their doctoral students Thomas Buchner and Toshihiko Fukushima are the co-first authors. They have reported their animal-inspired musculoskeletal robotic leg in Nature Communications  .

When robotic legs have to hold a certain position for a long time, a lot of current flows through the DC motor that drives them (left). Over time, energy is lost in the form of heat. In contrast, the artificial muscles (right), which work on the principle of electrostatics and are efficient, remain cold, because no current flows through them under a constant load. (Image: Thomas Buchner / ETH Zurich und Toshihiko Fukushima / MPI-IS)

Here is an exclusive Tech Briefs interview, edited for length and clarity, with Katzschmann, Fukushima, and Buchner.

Tech Briefs: What was the biggest technical challenge you faced while developing this robotic leg?

Katzschmann: We need to apply this to our robotic system, but also not in a laboratory environment. So, we need to overcome lots of challenging engineering. For example, in the laboratory you always have the actuator in a vertical position, and it's always straight; there's no other motion disturbing it. Also, maybe it’s just one set of pouches. Whereas, for the leg, because we needed more force, we would use several layers that would then kind of interfere with each other.

And all of this led to new things about how to have this in a more chaotic environment. And how to apply it to a system outside of an actuator testing setup.

Because, originally when we got started, we got the leg to just be able to stand up but not lift off the ground. Within six months, we got to a point where we could get it to stand up; we could have one big muscle to do it, but it wasn't at all at the level that we could get it to jump. There was lots of engineering, and we originally started in April 2021. So, it took quite some time to get to this point.

Tech Briefs: How did this project come about? What was the catalyst for your work?

Katzschmann: The research came a bit more from the direction of building soft robots and soft muscles. What we really wanted to do was go toward applying the insights from soft robotic technologies toward skeletal systems that have inherently the capability to overcome gravitational forces. Any animal in nature that has a skeleton can kind of, through evolution, get out of the water onto land. So, natural progression was to say, ‘Let's build a robot with muscles,’ but ideally use a muscle technology that has the potential to be efficient and to be all onboard and not, let's say, use external compressor units.

So, it was very logical to say that, together with Max Planck and ETH, ‘OK, we are the systems people; we build the robots. Max Planck will make the robotic materials. We'll make the film units, the HASEL units. And, so, that's why we teamed up and said, ‘OK, let's bring this together and make a nice synergy and build a robot with these muscles.’

Tech Briefs: Can you explain in simple terms how everything works?

Katzschmann: The actuators are oil-filled plastic bags. About half of each bag is coated on either side with an electrode made of a conductive material. When we apply a voltage to the electrodes, they are attracted to each other. As one increases the voltage, the electrodes come closer and push the oil in the bag to one side, making the bag shorter.

Pairs of these actuators attached to a skeleton result in the same paired muscle movements as in living creatures: as one muscle shortens, its counterpart lengthens. We use a computer code that communicates with high-voltage amplifiers to control which actuators contract, and which extend.

Tech Briefs: Do you have plans for further research?

Katzschmann: We want to push toward this direction; we want to build quadropads and bipads that will show this ability, but also we want to go toward other parts of humanoid-inspired design. We really want to show that we can build robotic creatures that have the ability to use muscles instead of motors. And we want to ensure that they can be more adaptive in a rough terrain. Particularly, if you think of rescue scenarios, you're thinking of really unstructured scenarios where we would have to go over rough terrains. We have to be able to sustain harsh conditions. That's really the long-term vision of our research.

That's also what's driving both Toshi and Thomas in their research: To say, ‘Can we build robots that can not just be working in the lab or outside on the grass, but can actually move further toward being in a cave?’

Fukushima: I think, for this, our main challenge will be how to make all the technologies onboard. So far, our robot is powered externally, with the high-voltage amplifier. But, to be a rescue robot, everything needs to be onboard and all the systems need to be integrated.

Katzschmann: We've been working on one direction where we are looking into making the muscles with a different material system that helps to then use other electronics, which then, in return, helps to put it all onboard. Also, thinking about the other components besides muscles, we need tendons, we need ligaments, we need new ways of thinking about joints that can be more bio-inspired, almost biomimetic. In all of these things we can learn from nature and apply them to our robots.

Tech Briefs: Is there anything else you'd like to add that I didn't touch upon?

Katzschmann: Maybe one thing that you didn't touch upon, and I think would probably be important, is the normal impression of a robot is that you have motors in the joints. And maybe people are not aware of this. When they look at a robot, they just think of a robot. But, if you think really carefully about what they all have in common, they put an electromagnetic motor with a gearbox into all of their joints. And we believe this is extremely unsustainable and not always the right direction to take for robots that should be human-centric.

So, robots that should be in an environment and doing things where also humans are in, because these metal machines, if they fall on you, it could hurt you quite badly. So, that's why we believe building robots with muscles and using other material technologies has a huge potential of rethinking the embodiment — when you think of embodied AI — the embodiment parts being made of a muscle-based technology could be the paradigm shifts away from these motor-driven robotics.