Helpful Robot Friends
Thousands of infants born each year with sensorimotor, social, and cognitive impairments can benefit from early motion interventions, such as body-weight-supported locomotion, that are designed to encourage motor exploration and practice. Naomi Fitter, assistant professor of robotics, thinks robots can act as an extension of physical therapist care. Supported by a National Science Foundation grant, she is developing novel hardware and algorithms for robot-mediated physical therapy interventions that can be tailored to individual children.
“Introducing a robot to help a child go through a physical therapy routine can help physical therapists work with more patients at once, help human therapists provide new types of intervention that don’t currently exist, and fill in gaps between physical therapy appointments,” Fitter said.
For a recent study, Fitter, who works closely with child development specialists, occupational therapists, and kinesiology experts, programmed a robot to model leg-extension kicks in front of 6-month-old infants. She found the infants in the study imitated the leg motion more often when the robot engaged them with light and noise rewards. Her NSF grant will fund additional studies in clinical and home-based settings.
Fitter knows that for her physical therapy robot to succeed, it will need to have a pleasant bedside manner. Among her efforts to make human-robot interactions smoother and more playful are a robot programmed to deliver the perfect high-five; a robot that roams people’s desk and reminds them to take work-day breaks; and Jon the Robot, a joke-telling robot inspired by Fitter’s background as a standup comedian.
“It’s important that as these systems emerge in spaces where we’re living and interacting and spending our day-to-day lives, that they’re intelligent, ethical, and fun to interact with,” Fitter said.
Fitter discusses her research on human-robot interactions in season 9, episode 6 of “Engineering Out Loud,” the podcast of the College of Engineering at Oregon State University. Subscribe on Spotify or Apple Podcasts, or listen online at https://engineering.oregonstate.edu/outloud/.
When Your Colleagues Have Microchips (LOL)
Bill Smart, professor of mechanical engineering and robotics and associate director of the Collaborative Robotics and Intelligent Systems Institute at Oregon State, wants to simplify interactions between robots and people that work together. So, Smart — recently named an Amazon Scholar — made the most of the colossal size of Amazon and its dozens of fulfillment centers, where more than 200,000 robots efficiently handle items alongside humans.
As an academic, Smart can perhaps run a few robots for a few days at a time and conduct what amount to proof-of-concept studies. But working at Amazon allows him the freedom to think on a grander scale — over the course of months or years — and to ask questions that are usually beyond the limitations of academic research. Of particular interest to Smart are the ways in which a robot signals its intentions of where it plans to go next.
The robots he currently works with have a set of indicator lights, similar to a car, that show their intent, and their underlying safety systems will cause them to slow or stop when they encounter humans. His ultimate goal: Make the robots “invisible in use,” so employees don’t have to think about them any more than they think about the actions of their human colleagues.
Bracing For the Next Surge
With intensive care units overfilled, health care workers stretched thin, and critical equipment in short supply, managing a COVID19 surge requires making a lot of tough choices, and quickly. Researchers at Oregon State, with colleagues at North Carolina State University, are working to understand the decision-making processes of hospital systems before, during, and after the pandemic to help hospitals better manage capacity for future pandemics and other mass casualty events.
“Our efforts should lead to better understanding of how hospital systems can mount an effective response to these major disturbances,” said Joseph Agor, assistant professor of industrial engineering at Oregon State, who is leading the effort.
Supported by a Rapid Response Research grant from the National Science Foundation, the group has been collecting and analyzing data from two hospital systems — Samaritan Health Services in Oregon and MedStar Health System in the Washington, D.C., metro region — to document operational changes, policies, and practices deployed to manage the surge of patients.
Agor talks about managing the COVID-19 surge in episode 5 of season 10 of “Engineering Out Loud,” the podcast of the College of Engineering at Oregon State University. Subscribe on Spotify or Apple Podcasts, or listen online at https://engineering.oregonstate.edu/outloud/.
Farm-Fresh Produce, Raised by Robots
Joe Davidson, assistant professor of robotics and a member of the Collaborative Robotics and Intelligent Systems Institute at Oregon State, is working to develop robotic systems for agriculture. His interest in robotics began at Washington State University, where he earned his doctorate working on a project to mechanize apple harvesting.
Teaching a robot to pick an apple isn’t so easy, according to Davidson.
“That’s because the agricultural environment is so highly unstructured,” Davidson said. “It’s not like an automotive factory, where you have a very controlled environment, and parts can arrive at the exact same place and the exact same time, over and over again. Every tree is different.” Davidson and graduate student Alejandro Velasquez train their robotic arm to perfect its picking motion on a synthetic apple that mimics the size and weight of the real fruit. The force required to pull the “apple” from its stem is replicated by external springs and internal magnets.
In addition to his work on harvesting, Davidson is working to develop robotic systems for tree pruning, along with Cindy Grimm, professor of mechanical engineering and an expert in robotic grasping and manipulation. Both projects are funded by the U.S. Department of Agriculture.
The scope of Davidson’s research has recently branched out into precision agriculture — a data-intensive farming paradigm that targets inputs (e.g., nutrients and pesticides) to individual plants — with a project funded by the Washington Tree Fruit Research Commission. Davidson envisions smart systems that will determine the specific needs of every single tree in an orchard. “If we have a system that makes the right decision for each tree, we can achieve optimal yields by making targeted applications,” he said.
Learn more about the benefits of robots in agriculture in season 9, episode 5 of “Engineering Out Loud,” the podcast of the College of Engineering at Oregon State University. Subscribe on Spotify or Apple Podcasts, or listen online at https://engineering.oregonstate.edu/outloud/.
Just in the United States, tapping the power of ocean waves could represent 2.64 trillion kilowatt-hours in theoretical annual energy potential — equivalent to twothirds of the nation’s electricity generation. Working to maximize this potential is Bryony DuPont, Boeing Professor of Mechanical Engineering Design. With funding primarily from the Department of Energy, DuPont develops computer simulations to help wave energy converter manufacturers design devices that wring the most electricity out of every wave.
“As you start designing these novel devices, there’s a really wide design space,” DuPont said. “And while people come up with really creative solutions, it’s not clear which yield the best performance.”
Beyond just performance, DuPont’s models attempt to optimize other considerations such as cost, reliability, manufacturing, and deployment. Finally, wave dynamics can vary considerably across locations. To account for this, DuPont is investigating three different wave energy converter archetypes in eight different spots around the globe to assess the ideal match between archetype and location.
“The energy is out there, just waiting to be harnessed,” DuPont said. “We want to help the DOE and our industry partners capture as much of it as they can.”
Faculty Advances in DOE Solar Desalination Contest
Espiku, a company with strong ties to Oregon State University, has advanced to the semifinals of the Department of Energy’s American Made Challenge on Solar Desalination.
According to Bahman Abassi, company founder and assistant professor of energy systems engineering at Oregon State’s Cascades campus in Bend, the project team aims to design and build a pilot plant to recover clean water from wastewater produced by oil and gas extraction, with zero liquid discharge.
“As more communities face water shortages due to climate change, these new systems could expand the use of nontraditional water sources and ease the burden,” said Kelly Speakes-Backman, acting assistant secretary for the DOE’s Office of Energy Efficiency and Renewable Energy.
Abassi’s company, which was selected to the group of eight semifinalists from 162 initial applicants, has received $250,000 in cash and a $100,000 voucher to continue its work. Advancing to the next round would yield $750,000 and another $100,000 voucher, with the ultimate winner taking home $1 million.
Great Strides for Cassie the Robot
Cassie, the bipedal robot being developed by Oregon State spinoff Agility Robotics, racked up two impressive feats recently: running a 5K footrace and climbing a set of stairs without the benefit of cameras or other sensors. Instead, Cassie relies solely on proprioception, or body awareness, to guide it — a first in bipedal robotics.
These demonstrations represent the culmination of two years of research on “sim-to-real” reinforcement learning in the Dynamic Robotics Laboratory, said Yesh Godse, a senior in computer science and applied mathematics who works in the lab. The use of computer simulation allows robots to learn how to navigate different types of terrain without damaging itself in repeated falls and crashes. This learning can then be transferred to the hardware to be tested in the real world.
“Things have gone quite well for our lab,” Godse said. “We’ve been established as having the most state-of-the-art research for sim-to-real reinforcement learning on Cassie, and perhaps legged locomotion in general.
Sarah Oman joins the faculty this fall as a senior instructor. She earned her doctorate in mechanical engineering in 2012 from Oregon State University. Prior to joining the faculty, she was the senior design capstone coordinator for mechanical engineering at Northern Arizona University. Her academic areas of interest focus on design theory and inclusive design.