Harvard University: 3D Knitting for Pneumatic Soft Robotics

December 04, 2023


We spoke with Vanessa Sanchez, a fashion designer turned engineer. She started at the Fashion Institute of Technology, pursued her BS in Fiber Science at Cornell and previously completed her PhD in Harvard SEAS and the Wyss Institute as part of the Harvard Microrobotics Laboratory.


In this interview, we chat about how the Kniterate machine that Harvard acquired was used to create all the samples for her academic paper 3D Knitting for Pneumatic Soft Robotics. Vanessa and her team studied the mechanics of knit structures to develop 3D-knit pneumatic actuators and devices with different motions and applications.

The workflow of creating, testing, and recycling a sample.

Why was Kniterate important for your research project?

The Kniterate was important for our team to study the mechanical properties of knits for use in 3D knitted pneumatic actuators. Something really good about the machine is how uniformly it could knit the samples. This consistency was key for us to actually pursue this research direction. If we don’t have uniform samples, what are we even testing – are our results changing because of the variable of interest or something else? With manual machines, knits aren’t always consistently uniform, especially because of human-operator differences and errors.

Manual knitting of complex structures and 3D devices can also be incredibly tedious and technically difficult to produce. It would take so much time, thought, and effort to go through the entire process for each sample (lowering the yield with inevitable mistakes). With the Kniterate machine, once you do it once, you can just keep on printing them off. Repeatability and efficiency were key factors. This made it possible to knit many structures of interest as well as several of these full actuators to compare them with each other and extract meaningful data.

The 3D Knit Inspection Robot was created by combining different types of stitches.

Why did you choose Kniterate over more established knitting machine manufacturers?

Harvard did not have a knitting machine yet, and there’s always risk associated with adding a new technology. We identified that the Kniterate was a good candidate to start getting knitting into the university and try things out.

The Kniterate doesn’t take up as much space as an industrial machine. Especially in research labs, where space is always a major constraint, the size of an industrial machine would be an issue. The Kniterate’s price point allowed us to minimize and take on the risk with starting a new research direction. When starting the work, we only had hypotheses, so there still were looming questions like; “Will this project work? Can we make these actuators? Can we knit some of these materials of interest?” These factors in cost and size really helped us to initialize this work, leading to a successful research program.

The 4×4 garter stitch structure forces the tube to bend when inflated.

What is the part that you found more surprising using the machine?

I think the learning curve and machine configuration process have similarities to those seen in 3D printing. The Kniterate requires you to spend time with the machine testing things out, but once we got to that point, it really opened up so many more possibilities beyond what we thought possible. What you put in, you get out.

We were doing a lot of weirder stuff. We made full 3D knitted pneumatic actuators using half gauging of double-bed structures, added in conductive yarn and monofilaments, tried directed plating, and things like that. The machine is very robust, so you can try out a lot. Some industrial machines are so precise that you’re more likely to break something when running more “experimental” programs. We push to the limit and see. Sometimes, we have tangled yarns to remove, but the machine could handle things and is up to run the next program. Also, it’s pretty easy to replace the needles when required.

Image taken from the Supplementary Information of the paper.

What do you find exciting about the future of digital knitting technology?

I feel like knitting is going in the way that 3D printing is, and it’s becoming more accessible to people. It’s just such a great on-demand additive manufacturing technology, where we can do so many things that are complementary to traditional 3D printing methods. We can make extremely large and open structures, and incorporate many diverse materials into monolithic networks. We can make lots of new devices, but also form everyday clothing on the same machine.

We’re testing things in a lab, but with digital knitting, it’s very promising that our new devices will be able to scale for manufacturing. We’re also excited that our findings can be broadly applicable and shared with other researchers who can use this understanding on other machinery.

I’m excited to see more people getting interested in knitting, mirroring the trend we’ve seen with 3D printing, where it has become a popular fabrication method both for professionals and hobbyists at home. I believe digital knitting is on a similar trajectory.

The 3D Knit X-Gripper was created by combining elastomer with PVA yarn, that shrinks when immersed in water to increase material stiffness.

Anything else you would like to add?

I don’t think this project would have gotten this far without Kniterate. Additionally, many people are interested in seeing the machine when it’s running. It has brought people together because it is in such a prominent place that people often ask, “What is that? What’s going on? How does it work? What’s happening?” I think it’s a nice tool in a research space or university, as everyone sees it in action, and it becomes a nice way to get people to start talking or collaborating.


More information about the research:


Tags: Showcase | Back to Blog