Matter Assembly Computation Lab

Background

I began my work at the Matter Assembly and Computation (MAC) Lab at CU Boulder after my senior year of high school. I have worked under the excellent mentorship of Professor Robert MacCurdy and PhD student Charles Wade. I worked at the MAC Lab over the summer and throughout the school year, beginning the work on my primary research project, increasing the performance of Pneumatic Network Actuators through multi-material 3D printing.

Pneumatic Network Actuators (PneuNets) are a series of connected hollow pockets that are inside a soft finger-like body that is usually made of silicone. When the air inside the pockets is pressurized, the body will begin to expand, resulting in a bending moment at the base of each pocket, causing the finger to curl. These actuators are excellent for applications where a robot needs a gentle touch, such as medical procedures and manipulating fragile objects. One of the main issues with these actuators is that their soft bodies have a hard time firmly grasping objects before undesired bending and slipping.

I aim to increase the performance of these actuators by leveraging the capabilities of multi-material 3D printing, allowing the creation of a rigid endoskeleton within the soft body of the actuator. Another issue is that most PneuNets are made via casting silicone into a mold, which is a very time and labor-intensive job. 3D printing technologies can help solve this issue by reducing manufacturing to a single hands-off step.

Process

For my project, I use a Stratasys J750 Ink Jet 3D printer. Ink jetting is a very interesting additive technology that few are aware of, and I am very fortunate to be able to work with it at the MAC Lab. Using an ink jet system allows me to use very rigid materials alongside very soft ones, reaching around 30A on the Shore hardness scale.

Using this fabrication process, I can quickly make my actuator and then test it on an inflation test rig designed by another member of the MAC Lab. The inflation test rig uses computer vision to read indicator dots on the actuator to model its curvature vs. input pressure. After this, it will inflate the actuator to different pressures and record the transmitted force at each pressure. This data is then compiled into a graph before being analyzed.

Media

Inkjet 3D printed PneuNet actuator

CAD model of the endoskeleton PneuNet design

Inflation test rig with computer vision tracking

Actuator testing — horizontal and vertical inflation cycles

Vertical inflation test — endoskeleton PneuNet actuator

Personal Development

This project has required me to learn a large number of new tools, including Abaqus FEA, MATLAB, Ink-Jet Additive Manufacturing, LaTeX, OpenVCAD, and MAC Lab's Robotic Inflation Test Rig. Along with these tools, my skills for presenting, collaborating, academic writing, academic reading, researching, data acquisition, data management, and descriptive speaking have all been practiced continuously. Through two weekly meetings and preparation for conferences and paper submissions, I find it very rare for there to be any time that I'm not growing as a researcher. I am incredibly grateful to have been given this opportunity, and I greatly look forward to the years of my research to come.

Results

Preliminary data show that my endoskeleton PneuNet design can outperform traditional PneuNets for reducing undesired bending while still maintaining similar in-plane bending efficiency. I was preparing a paper to present this information at the IEEE IROS conference, but unfortunately, the J750 started to have issues, and I wasn't able to consistently manufacture samples. Without a robust data set, my IROS paper wasn't submitted. However, I was selected to present my preliminary data at the National Conference on Undergraduate Research, which was a great way for me to share what I have done and learned so far.

Next Steps

I have been working on my paper over the summer and plan to collect data when I gain access to the working J750 in the fall. I plan to submit my paper to a different IEEE conference to present what I have found. There are countless ways to improve the design of PneuNets to fully leverage the capabilities of multi-material 3D printing. I plan to continue working on this project, increasing the functionality and capability of these actuators. Hopefully, I will have some exciting news about publications on my webpage soon!