Robot Melts Its Bones to Change How It Walks

Joints that are reconfigurable on the fly help this small robot avoid obstacles

Robot that melts its bones
Photo: Colorado State University

Different animals are optimized for different things, and this optimization is reflected in the structures of their bodies. It’s especially evident in the skeletons of animals designed to move around on land, where there’s a crazy diversity of limbs and joints and feet. There are some generalizable structures that tend to work well, like having hips and knees and ankles and feet, but if you look at the difference between the skeleton of an ostrich and the skeleton of an elephant, you’ll get a sense of just how much wiggle room there is.

Unfortunately for animals, optimization means that while they’re excellent at some things, they struggle with other things, because they’re not able to redesign and re-optimize their skeletons on the fly, because how on earth would that even work, right? But robots suffer no such biological constraints, and roboticists at Colorado State University have developed a small walking robot that can melt and solidify its bones on the fly to optimize its legs for different motions.

A traditional approach towards making multimodal robots has been to just zip-tie or staple or duct tape or whatever different modalities together into one robot, whether it’s legs and wings, or wheels and rotors, or any other combination you can think of. The issue with this has always been that you end up spending a lot of effort on a big chunk of your robot that may not be useful much of the time, which is an inefficient way of going about things. A more efficient approach is to make a robot reconfigurable in some way, so that it uses the same hardware for a diverse array of capabilities.

Jianguo Zhao, whose work on jumping robots we wrote about when he was a graduate student at Michigan State, is now an assistant professor at Colorado State. In a paper published in IEEE Robotics & Automation Letters, Zhao introduces a new kind of small reconfigurable robot that relies on a plastic structure that can be selectively melted and re-hardened on the fly to change joint configurations and resulting motion. The researchers call these “shape morphing joints,” but I feel like it’s more understandable to think of them as bones that morph into joints and back again. 

We propose a new reconfiguration strategy by introducing into a mechanism with shape morphing joints (SMJs), which can be either soft for a compliant joint or rigid for a structure. With several SMJs in a mechanism, we can change its functionality by strategically softening and solidifying appropriate SMJs, without altering the underlying mechanical design after it is fabricated. Further, we use the same motor that drives the mechanism to transform it to other configurations. Potentially, reconfigurable mechanisms with SMJs can enable adaptive robots that can adjust their sizes, shapes, or functions to fulfill multiple tasks in different environments.

It takes just over 10 seconds to soften the PLA that the robot’s joints are made of using a wrap of wire that heats up when voltage is applied it. A silicone cover keeps the PLA hold its shape while it’s compliant. Depending on which of its joints are softened, the robot in the video above can switch between a rigid structure, a four-bar linkage, and a five-bar linkage, and you get all this with just two joints—it’s easy to add more. 

This particular robot is about as simple as a walking robot can be while still being able to take advantage of these new joints: it’s mostly 3D printed, and only uses two actuators in total to move all four legs. The thing to be more excited about is using this system on robots that scale down, where improving capabilities without adding cost or mass or complexity will be a big advantage.

For more details, we spoke with Jianguo Zhao, who directs the Adaptive Robotics Lab at Colorado State, via email.

IEEE Spectrum: You say in the paper that you’re interested in leveraging the same body parts for different functions in the way that many animals do, but animals can’t reconfigure the structures of their bodies in the way that your robots can. Why did you decide to use this approach, rather than modeling your robot more closely on the structure of an animal?

Jianguo Zhao: Animals have numerous muscles distributed in their body parts, and they can selectively actuate those muscles to achieve different functions. Although it is possible to mimic such distributed actuations using artificial muscles, the associated modeling and control will be extremely difficult. To circumvent such a problem, we choose to reconfigure the mechanical structure of a robot, which can also achieve multiple functional found in animals (different motions for walking, jumping, or swimming, for example).

What kinds of things will robots using this technology be able to do that animals can’t do?
For our reconfiguration method, robots can morph to another configuration and hold that new configuration without any energy consumption. Even though an animal can achieve new configurations, it’s impossible for it to hold that configuration without any energy consumption. For example, a bird can span its wings during gliding, but it needs to actively actuate its muscle to maintain the wing shape.

Can you give a few examples of other kinds of robots that could benefit from this technique?

We envision our technique can be applied to many other robots, especially robots for locomotion. For locomotion in the same medium (on land, in the water, or in the air), robots can morph their shape to travel more efficiently. For example, walking robots can change their gaits to accommodate different conditions. Flying robots can morph the shape of their wings to cope with diverse aerodynamic situations. Our technique can also potentially endow robots to travel in different mediums to achieve multimodal locomotion. For instance, an amphibious robot can morph its forelimbs to be more suitable for walking on land or for swimming in the water.

What applications will robots like these be ideal for, and how will their ability to morph help them in those applications?

Our morphing technique is ideal for robots that are small but need to perform different tasks or adapt to different environments. Those robots can be used for a wide range of applications including environmental monitoring, military surveillance, as well as search and rescue in disaster areas. With a small size, we expect those robots to be low cost so that we can build many of them economically. In this case, we can potentially deploy them for a specific application automatically to form mobile sensor networks and work collaboratively to accomplish given tasks.

What are you working on next?

We are working on two aspects for the morphing concept. First, we try to extend the proposed method from generating planar trajectories to three dimensional trajectories. In this case, we hope to accomplish robots with multimode locomotion (e.g., swimming, walking, or flying). One particular interesting question is how can we achieve a set of different motions or morphologies by properly designing the mechanical mechanism. Second, we are trying to improve the performance of the technique, such as reducing the required time for the morphing process and improve the reliability of the shape morphing joints. This can potentially lead to real-time shape morphing or reconfiguration in dynamic environments.

“An Adaptive Walking Robot With Reconfigurable Mechanisms Using Shape Morphing Joints,” by Jiefeng Sun and Jianguo Zhao from Colorado State University, is published in IEEE Robotics and Automation Letters.

[ Adaptive Robotics Lab ]

Robotics News

Biweekly newsletter on advances and news in robotics, automation, control systems, interviews with leading roboticists, and more.

About the Automaton blog

IEEE Spectrum’s award-winning robotics blog, featuring news, articles, and videos on robots, humanoids, automation, artificial intelligence, and more.