Elizabeth (Ellie) Andreyka

UT Austin’s HeRo Lab, I worked on developing a method to provide haptic guidance in surgical training using spatial cues from a vibrotactile sleeve.

I was responsible for designing, building, conducting troubleshooting, and integrating mechatronic systems, software, and virtual environments for the project. The virtual reality environment was developed in Unity, utilizing C#.

I successfully executed a 22 person human user study that received approval from UT’s Institutional Review Board (IRB). Post-study, I conduced a comprehensive data analysis using MATLAB.

For the visually impaired, in an increasingly digital world, access to written information has become more difficult. Because of this, research has focused on making refreshable Braille displays cost friendly and effective for linguistic communication and learning.

On a team of two other graduate students, we created two Braille displays: one designed in a previous study using solenoids [1], and the other a novel cost-friendly vibrotactile sleeve with six vibration motors. Participants were tasked with single letter Braille reading comprehension receiving guidance from the two devices compared to a standard Braille training board. To assess the effectiveness of each method, the time it took each participant to guess a letter (comprehension time) and accuracy of the guess were recorded. The comprehension time using the solenoid device was 43% faster than the traditional Braille print, and the vibrotactile sleeve was 11% faster than the traditional Braille print. Both the solenoid and vibration actuator devices resulted in 79.5% accuracy for the Braille readings, compared to 75% for the standard Braille print.

[1] G. C. Bettelani, G. Averta, M. G. Catalano, B. Leporini and M. Bianchi, "Design and Validation of the Readable Device: A Single-Cell Electromagnetic Refreshable Braille Display," in IEEE Transactions on Haptics, vol. 13, no. 1, pp. 239-245, 1 Jan.-March 2020, doi: 10.1109/TOH.2020.2970929.

Impact Studio is an organization affiliated with Oregon State University created to respond to trends in higher education and better meet the needs of students.

As a design intern at Impact Studio, I researched, designed, and developed program initiatives put forth by students and faculty at Oregon State to explore their prospective implementation in the University. Our multi-disciplinary team of students, faculty, and staff worked both collaboratively and individually to push these initiatives forward, and took a human-centered design approach to developing each program.

I most heavily worked on the Engineering and Design for Society (EDS) initiative which is primarily centered on an academic program (certificate, minor, and major) that exists at the intersection of engineering, design, business, and liberal arts with the goal of creating a multi-disciplinary design experience for students at OSU.

I worked on various parts of the project over three years: performing research on similar programs at other universities, doing student research and focus group facilitation, and writing project proposals. I worked with the faculty of the college of engineering, business, liberal arts and university administrators to move this project forward. The program went into pilot at Oregon State in Fall 2021.

CREATIVE NON-FICTION:

I was a writing minor as an undergraduate, and did my Honors College thesis through the school of Writing, Literature, and Film. The project was a series of creative non-fiction essays on the experience of being an undergraduate woman in STEM, and explored various creative-nonfiction modalities to tell stories on the subject. The thesis can be found here. It won the Honors College Outstanding Thesis in the Humanities/Social Sciences for 2021.

Letterpress printing is pretty neat! Some of my letter-press printing work can be seen to the left.

For my undergraduate senior capstone project, I worked with with two other students to do material testing for Celilo Cycles bike frames. Celilo Cycles bicycle frames are made out of wood, which has advantages in shock absorption and sustainability, but the lifetime of the frames was significantly less than that of competing products. To increase durability and lifetime, Celilo Cycles wanted their bikes to be reinforced with basalt fiber, which is an effective, sustainable, and cost-efficient reinforcement material, similar to carbon fiber and fiberglass.

Our job was to create variations in fiber weave and epoxy type and test them under multiple loading conditions with an Instron Machine to determine mechanical properties of the frame (ultimate tensile strength, young’s modulus, and ultimate compressive strength). These values, as well as the usability, manufacturability, and flexibility of the fiber, were used to inform Celilo Cycles on the best basalt fiber material combination to use as reinforcement.

As a research fellow at Los Alamos National Lab with the Los Alamos Summer Dynamics school, I researched neutron radiation detecting Optically Simulated Luminesce (OSL) fibers. Variations in the OSL fibers were tested in Monte Carlo N-particle (MCNP) software to determine the optimal configuration for neutron capture and detector efficiency.

Variations in the fiber core size, moderator size, and moderator material were tested under various neutron radiation sources, including a neutron point source and AmLi and Californium source spectrums. To do this, a Python script was written to create new simulation files to run on MCNP. Given this data, ideal moderator material and size, as well as the fiber core size, were determined for applications in a wearable textile format.

These fibers are different from personal neutron detecting radiation dosimeters in that they do not require power, need to be hard-wired, rely on the direction of radiation dose for accuracy, have limited lifetimes or need to be recalibrated often, have significant time delay between recorded dose and feedback, and they are reusable, provide feedback as often as is needed, can store dose data, and can potentially be used for spatial dose information.

The ultimate goal of this research is for the fibers to be used in a wearable textile format, such as a glove for a glove box or something that covered more body mass, like a suit. The positives of using a woven textile format is the ability to pinpoint where in the textile there is the most radiation. With this information, a wall or building could be wrapped in the textile, with different regions showing where the most radiation is.

Our resulting published work on this research can be found here, or under the title Simulated Optimization of Neutron Detection in Optically Stimulated Luminescence Fibers.

[1] Gong, Z.; Xiang, Z.; OuYang, X.; Zhang, J.; Lau, N.; Zhou, J.; Chan, C.C. Wearable Fiber Optic Technology Based on Smart Textile: A Review. Materials 2019, 12, 3311

I worked for three years on the OSU Mars Rover Robotics team to develop a Mars rover for competition. I worked on designing a mechatronic system to dig up soil, test it, keep it secure, and deliver the sample autonomously. I also worked on designing the electronics box for the rover and to fit all of the necessary electrical components inside utilizing SolidWorks.

3D printers have come a long way, but there are still a lot of barriers to its use in mass production. Desktop 3D printers usually use a chemical adhesive (glue, certain coverings, etc.) to adhere a print to a print bed. This adhesive tends to not be strong enough to hold the first few layers of a print to the bed. Because of this, the first layers of a print need to be highly monitored and constantly adjusted in order to get the print to stick in just the right way. If a print’s first few layers are off, the entire print becomes warped.

The high level of monitoring and human intervention required in the first part of a print is a huge barrier to utilizing 3D printing in mass production.

Oak Ridge National Lab’s Manufacturing Demonstration Facility developed a print bed that mechanically captivates a print instead of chemically. The print bed has grooves which the first layer of the print fall into and then secure the print into place. As a robotics intern at Oak Ridge, I was asked to use this bed in a conveyor belt 3D printer design. The printer would print as normal on this print bed, and when it was done it would rotate the bed releasing the print and go on printing the same thing again until it ran out of filament/material. This design would make mass production of 3D printed parts much more feasible, as there is little to no monitoring needed to print a massive amount of items.

Lack of access to clean drinking water is a leading causes of death in developing countries. The most common method of cleaning water in these areas is to boil it over a wood-burning stove. However, producing enough potable drinking water for every person in a developing country through the traditional wood-burning stove method would require an unsustainable amount of fuel and produce large quantities of harmful emissions. Gathering wood for fuel can also cause deforestation, namely in sub-Saharan Africa where many forests have been cleared out. The loss of these forests has meant habitat loss, watershed damage, desertification, and climate change which all further push the issue of regional food scarcity. Also, with this deforestation, women and children have to travel farther and farther from their homes to get fuel for boiling water, putting them at a higher risk for sexual violence and kidnapping when gathering fuel. 

The InStove Water Purifier is a fuel efficient water pasteurization system that uses around 90% less wood than traditional wood burning stoves, or about a pencil’s weight of wood for one liter of water. The Water Purifier can produce upwards of 4,500 liters of drinking water a day with minimal financial and health cost to the operator and the environment. The massive quantity of water it can produce makes it ideal for larger institutions like hospitals or hostels. I worked on the micro-biology testing of the system, aided in a few iterations of re-design, and eventually field tested the system at a girl’s hostel in Mbale, Uganda.

Field testing involved observing user interaction with the system and interviewing users to strengthen the system’s design. These interviews significantly shaped the next design iteration. For example, a lot of money was spent on a heat exchanger which turns the hot pasteurized water to room temperature before leaving the system. However, most people we interviewed boil their water before drinking it, and said they didn’t trust that the cold water had been heated and cleaned. Taking this component out of the design would not only make it a more desirable product for the end user, but would also simplify the design and make it less expensive to distribute.

Field testing clearly demonstrated how important research and human centered design is in developing and designing a product, and how important it is to involve the end user in each step of the design process. After field testing, the feedback was used to simplify and adjust the system’s design, reduce turbidity, and develop easy and universal user manuals for the system’s operators. 

Here is a link to an accepted conference proceeding about the development and testing of the system.

Worked on a team to re-design the toes on the Cassie Bi-Pedal Robot. The goal of this project was to ideate a bunch of designs for better traction and balance, and narrow in on a design that is better suited for longer walks with perturbations in the ground such as a trail walk. Designs were ideated in SolidWorks, prototyped, tested for impact distribution, and the data was analyzed in MATLAB. A example ideation can be seen to the right.

PROBLEM:

It is very hard to mimic human muscles in robotic form. There are so many fibers, signals, and webs of tissue all working together to get your body to move. For robots modeled after mammals (robots with arms and legs and human-like joints) muscles are essential. Muscles support mechanically fragile joints, and allow the robot to be more adaptive to their environment — these are characteristics needed for robotic prosthetic development or human/mammal-modeled robots.

THE SOLUTION:

As an REU intern at Portland State University, I worked to develop flexible actuators that act similarly to human muscles. The design of these ‘muscles’ included a tube of strong and flexible material called ‘Festo’ with two end caps secured on the ends. Air would fill these tubes, and the pressure in the tube would be recorded over time. This data was compared to muscle contraction data, and the end caps were re-designed until the muscle contraction data most closely matched the internal pressure of the tube over time.