New one-of-a-kind artificial tissue is three times stronger than natural muscle (and could be used to create robots that move and work independently).
Advances in computer science and engineering are making robots smarter and more useful to humans than ever before. Scientists and startups have innovated to make robots better at understanding human speech, for example, and faster on the assembly line.
Now, they’re getting stronger: A research team from Columbia University’s Creative Machines lab has developed the first artificial tissue that can function in robots just like normal muscle.
The 3D printed material – which can lift up to 1000 times its own weight – is the only “soft actuator” capable of handling high levels of stress and strain without help from external compressors or high voltage equipment.
The soft actuator is a game-changing breakthrough for the expanding field of soft robotics. By creating a synthetic muscle that mimics normal biological processes, the researchers say they’ve overcome “one of the final barriers to making lifelike robots.”
The field of “soft robotics” is focused on developing flexible materials capable of extending, contracting, or bending in response to simple control inputs. By developing soft or “deformable” structures for robotic systems, researchers can help make robots more useful for grasping and manipulating objects (even delicate ones) or interacting with living organisms.
But when it comes to driving breakthroughs in soft robotics, the field’s multidisciplinary nature – spanning several different subsectors of engineering and materials science – poses challenges.
Compared to their peers, soft robotics researchers have much less peer-reviewed work to start from: Research into the technology-aided fabrication of unconventional materials (such as 3D printing of synthetics for medical devices) is growing, but much is still exploratory. Defining standards for which soft materials are most useful in robotics will demand much further study.
Soft robotics research also faces the challenge of piggybacking or improving on innovations in “hard” robotics, where (as Columbia Engineering professor Hod Lipson points out) there is a lot of work yet to be done:
“We’ve been making great strides toward making robots’ minds, but robot bodies are still primitive,” said Lipson, who led the study. “This is a big piece of the puzzle.”
The so-called puzzle piece developed by Lipson and his team is a 3D-printed material capable of pushing, pulling, twisting, and lifting weight in response to low-power stimuli. The low-power requirement marks a key difference between the Columbia innovation and existing soft actuator technologies, most of which can only expand when air or liquid is supplied to them. That process usually requires large machines (and thus can’t be miniaturized to life-size-robot scale).
When activated by the electric current of a thin resistive wire, the Columbia team’s soft actuator expands and contracts to execute tasks – such as gripping an egg or lifting a bottle, as the video below shows. Over a series of experiments, the video also shows how the actuator is capable of behaving like a human bicep, and of substituting for an electrical motor in a manufactured robot.
When electrically heated via computer controls to 80 degrees Celsius, the actuator was capable of expanding up to 900%. Assessments also showed that the new material has a “strain density” (expansion per gram) that is 15 times larger – aka three times stronger – than that of natural human muscle.
The actuator was created from a form of silicon rubber laced with microscopic ethanol particles; the solution was engineered to be highly elastic but also easy to fabricate at a low density, which is necessary for low-voltage activation. (The research notes that the solution is also environmentally safe and inexpensive.)
The combination of silicon and ethanol gave the actuator the high-strain and high-stress characteristics that set it apart from any similar developments to date. The researchers also say that the actuator can be shaped and reshaped a thousand ways – allowing it to perform motion tasks in almost any design.
“Our soft functional material may serve as robust soft muscle, possibly revolutionizing the way that soft robotic solutions are engineered today,” said Aslan Miriyev, a a postdoctoral researcher involved in the study. “It’s the closest artificial material equivalent we have to a natural muscle.”
As the most flexible, inexpensive, and effective artificial tissue developed for robotics purposes to date, the Columbia research is primed for commercial robotics applications. (The study was funded by Columbia and a grant for 3D-printed robotics from the Israeli Ministry of Defense.)
Before then, the team will continue building on the work done so far: Plans include incorporating conductive materials to replace the embedded wire, accelerating the material’s response time, and making it longer lasting.
In their most futuristic vision, the Columbia team sees their work helping robots become not just better, faster, or stronger but also more fully “humanesque”: Long-term, they hope to involve artificial intelligence to auto-control the soft actuator muscle, which may be a last milestone towards replicating natural, lifelike motion in the robots of the future.
The original study “Soft Material for Soft Actuators” was published in the journal Nature Communications in September 2017. Full information is available here.