From Exploration to Innovation: The Impact of SURF Grants
Media Inquiries
Researchers boldly explore uncharted territories shedding light on the unknown to advance human understanding. Their pursuit of knowledge fuels innovation like a river nourishes the land it flows through, leaving behind growth and blossoming possibilities.
Undergraduates at Carnegie Mellon University can join this pursuit through the Summer Undergraduate Research Fellowship(opens in new window) (SURF). A SURF grant enables young researchers to collaborate with esteemed faculty, build professional networks and receive financial support to focus solely on their project, paving the way for future career or academic opportunities.
Hyojae Park, Ben Gu, Tiffany Han, Irene Cui and Nikash Bhardwaj are just a few of the many CMU students who have enjoyed this transformative experience. From a soft robotic hand that helps amputees to a small-scale robot for search and rescue, from an educational robot that’s kid-friendly and affordable to an origami simulation aimed at developing a better medical device, the projects these students chose not only pushed the boundaries of technology but may potentially improve people’s lives.
Advancing prosthetics through robotic hand research
Led by Nancy Pollard(opens in new window), CMU’s Foam Robotics Lab(opens in new window) is developing a soft, robotic hand that offers greater precision and flexibility than traditional prosthetic hands. When Hyojae Park became interested in contributing to the research, Pollard worked together with the computer science major to secure a SURF grant that would enable him to participate.
“The challenge with prosthetic hands is their limited range of motion, which makes it difficult for users to perform everyday tasks like folding laundry or peeling an orange,” Park said.
The lab is developing a robotic hand using soft materials closer to the idea of human skin, which allows for greater flexibility compared to rigid metal. These soft materials are tendon-driven, mimicking the human hand's structure where tendons connect to joints and are controlled by motors. However, replicating the complexity of a human hand, which has 46 tendons and dozens of muscles, is extremely challenging, not to mention expensive.
Park's approach involved using an optimization technique to determine the best set of tendons for a given set of motions.
He created a model to represent the tendons and then ran computations at the Pittsburgh Supercomputing Center(opens in new window) to determine the best set of tendons needed to most closely approximate the action of a human hand.
“We identified a bunch of different tasks, from turning a doorknob to picking up small objects, and ran that information through an algorithm,” he said. With an average error of about 2.6 millimeters per joint, Park said the results were highly accurate.
“It shows that the model was able to learn which tendons were good for which motions,” he said. “That can save designers a tremendous amount of time, because it eliminates a lot of physical tests of trial and error.”
The SURF grant also enabled him to present his work at a robotics conference in Nancy, France, where he had the opportunity to connect with other like-minded researchers.
Park developed his passion for learning while reading extensively at his local library growing up. That drive has continued into his academic career, where he is excited about the vast array of courses available at CMU.
"There’s an infinite amount of learning that's available at CMU, and I think that's so incredible," he said. “Planning my schedule of classes is always somewhat heartbreaking, because there’s not enough time to study everything I want to learn about.”
Looking ahead, Park plans to pursue graduate studies in robotics and computer science with a focus on developing fundamental algorithms for various applications. His research is not only advancing the field of robotics but also has the potential to significantly improve the quality of life for amputees.
Innovating control systems in small-scale robots
Ben Gu has been exploring the control systems of small-scale, two-legged robots. Guided by Aaron Johnson(opens in new window), an associate professor of mechanical engineering at CMU, his work involves designing and optimizing walking robots that measure 20 centimeters tall and weigh about 400 grams.
A foundational class taught by Johnson during his sophomore year sparked his interest in how motion works in relation to other parts and the environment. His passion for control systems was further ignited during a course on dynamic systems and control taught by mechanical engineering professor Mark Bedillion(opens in new window).
"What fascinates me about control systems is their versatility. They can be applied almost anywhere,” he said. “Control systems manage everything from automated plant-watering and synchronized fountains to stabilizing drones and noise-canceling headphones, from intelligent lighting to house climate control. It’s not just about making things move, but making them move exactly as envisioned, balancing creatively with technical precision to bring ideas to life in powerful ways.”
Gu credited the flexibility of being able to choose from undergraduate and graduate courses for enabling him to explore a variety of fields and find his true passion in robotics and control systems.
“I believe you have a lot more career opportunities with this background. By learning control, you’re not limited to robotics. It would be a powerful tool in many engineering fields,” he said.
He mostly appreciated the opportunity through the SURF grant to dedicate 40 hours a week to his research project without the distractions of academic coursework.
"It feels so different to do research during the summer. It’s great because it allows you to envision what a life as a researcher would truly be like,” he said. “Working on my robot over the summer was one of the most exciting and eye-opening experiences I’ve had in college. Research at Carnegie Mellon is amazing. You can raise your hand and say ‘we should do this,’ and then go ahead and explore your ideas.”
Origami-inspired simulations for minimally invasive surgeries
Tiffany Han is a senior computer science(opens in new window) major and an explorer at heart. She’s using origami principles to simulate the actions of foldable robots, which has a variety of potential applications including medical device design as well as search and rescue.
“Origami folds are quite useful when it comes to, for example, a medical device that you ingest, because the GI tract starts out very narrow and then expands, or a search-and-rescue robot that needs to know what shape it should take to get through a certain-sized crack,” she said. “Regardless of the application, if you have a folded structure that you want to use, a computer simulation can help you optimize your design.”
Han's journey into this unique area of robotics began with a waning interest in math and a fascination with the unconventional.
"Math was too theoretical for me, so I slowly moved into computer science," she said. “I chose the origami-inspired simulation project because it’s something that is very new.”
Her project involves simulating foldable robots in a physics environment, which aligns well with her computer science background. She uses reinforcement learning to train the robots to perform specific tasks, such as throwing a ball into a designated area.
Han's work is entirely computer-based, but she often creates paper models to visualize the structures she is simulating. Her goal is to develop tools that can automatically design these structures, making the process more efficient.
Han values the supportive environment at CMU, where she said she has received guidance from Ph.D. students and professors.
“Research at CMU is very collaborative,” she said. “Maybe I don’t know as much about these fields as a master’s or a Ph.D. student, but everyone is very patient and takes my ideas very seriously.”
Han plans to pursue a fifth-year master's program at CMU to deepen her knowledge in artificial intelligence, reinforcement learning and robotics. Her research not only advances the field of soft robotics but also demonstrates the potential of origami-inspired designs in practical applications.
Developing alternatives to plastic for use in health care
As a junior majoring in chemical engineering, Irene Cui is conducting groundbreaking research on biodegradable and renewable protein-based elastomers. Translation: she’s developing sustainable alternatives to the traditional plastics used widely in products — from car parts to medical equipment to shoes — that pose significant environmental challenges due to their difficulty in breaking down.
Cui’s interest in this research began during her sophomore year, when she discovered a professor in the Department of Chemical Engineering(opens in new window) working with proteins in innovative ways.
"I had never really considered proteins as materials. I'd always just thought of them more as biomolecules, things that are part of how we function versus what we can do with them," she said. “This new perspective intrigued me.”
Protein-based elastomers are flexible, rubber-like materials derived from renewable biological resources, which could potentially replace petroleum-based products in items like toys and water bottles.
"By using renewable materials, my research aims to reduce plastic waste along with its environmental impact,” she said.
Cui's dedication to her research was further fueled by the opportunity to work full-time on her project over the summer, thanks to the support and resources available through the SURF grant.
"Being able to focus on my project full time over the summer gave me a lot more time to really work with it and make progress," she said.
Looking ahead, Irene is considering various career paths, including pursuing an integrated master's degree, a Ph.D. or even medical school. "I'm really kind of just going with the flow and seeing which opportunity arises and what works for me," she said.
Cui's research not only contributes to the development of sustainable materials but also highlights the importance of interdisciplinary approaches in tackling environmental issues.
Machine vision to enhance the development of educational robots
Nikash Bhardwaj is a senior computer science major who wants to attend graduate school to study robotics and artificial intelligence. He thought doing undergraduate research during the summer would be a good way to become a better, more independent researcher.
His journey into computer science began with a course taught by graduate student Austin Schick during Bhardwaj’s first year at CMU.
"Austin just taught the course so well that it made me want to pursue a career in computer science,” he said. “I saw how applicable and useful it was with everything that’s going on in the world with AI and robotics. It seemed like a really interesting field to get into.”
Bhardwaj's SURF project involved creating a system that allows robots to intelligently navigate their environments using a single camera and a deep neural network to generate depth maps. Recently developed deep learning-based depth estimation techniques open the possibility of replicating the capabilities of high-cost systems like Tesla's autonomous vehicles in a more affordable way.
“We can see our ideas come to life while the students gain valuable research experience.” — David Touretzky
“My adviser David Touretzky(opens in new window) and I believe doing this will enable younger students to become more involved in robotics and artificial intelligence,” he said. “Our goal is to get a robot to be able to intelligently navigate their environment, for example, to be smart enough to avoid falling off the edge of a table. Then, you could put those robots in classrooms with our software, and young students would not have to worry about things like the robot sensing the environment. They could concentrate on just having the robot do things in the environment.”
Touretzky highlighted the significance of this work.
“The world is fascinated with AI right now, and CMU is one of the leaders in K-12 AI education. One of the most effective ways to teach kids about AI is to introduce them to intelligent, AI-powered robots. The project Nikash is working on is part of an effort to make affordable, kid-friendly robots smarter than they've ever been before. The implications of this are truly exciting,” he said.
Bhardwaj's research is part of an independent study, and he continued working closely with his professor through the fall semester. They have made significant progress, including resolving issues with accurately mapping environments, putting them close to having a functional demo.
“CMU undergraduates are tremendously talented, so working with them is rewarding for professors,” Touretzky said. “We can see our ideas come to life while the students gain valuable research experience.”
