Brenna Argall’s passion for robots started with flesh and blood.
As an undergraduate mathematics major at Carnegie Mellon University, she attended a seminar where the professor conjured the image of nanoscale robots swimming through the bloodstream to chew up blood clots. The powerful depiction — which seemed culled from a Stanislaw Lem or Neal Stephenson novel — offered a tantalizing new research direction for Argall.
“It wasn't until starting my PhD that I learned that those nano-robots were only simulations — they didn't yet exist,” says Argall (mechanical engineering and electrical engineering/computer science). “Today we do have nanoscale robots, but 10-plus years ago the smallest were just microscale.” That seminar, though, encouraged her to pursue to a two-year stint as a research assistant in a neuroscience lab at the National Institutes of Health. Afterward, the Wisconsin native earned a PhD in robotics from Carnegie Mellon, where she was affiliated with its renowned Robotics Institute. Since then, Argall has pursued research at the intersection of artificial intelligence, rehabilitation robotics, and machine learning.
Robots loom large in the popular imagination, thanks to a century of portrayals in science fiction where they appear variously as friend, foe, or some ambiguous hybrid. Their rise is partly attributable to the hopes and fears associated with twentieth-century technological advances. Famous examples include Maria, the golden android from the 1927 film “Metropolis,” which inspired the mincingly brilliant C-3PO in “Star Wars.” More terrifying figures are Arnold Schwarzenegger’s “terminator” — a time-traveling robotic assassin — or the enslaved replicants in “Blade Runner” who revolt against their human creators.
Argall’s team at Northwestern and the Rehabilitation Institute of Chicago’s Sensory Motor Performance Program aims to bridge the machine-human divide, including by developing a commercially feasible “smart wheelchair.” The innovation, perhaps within five years, could dramatically improve mobility in patients with paralysis, including those stricken by amyotrophic lateral sclerosis, multiple sclerosis, Parkinson’s disease, and traumatic brain or spinal cord injury.
In her patient-centric research, Argall is driven by a central irony. “Often, the more severe a person’s motor impairment, the more assistance they require, and yet the less able they are to operate the very assistive machines created to provide this assistance,” she says. Her lab addresses this challenge through customized robotics autonomy and intelligence that shifts the burden of control from the patient.
This year, Argall’s research earned her the prestigious Faculty Early Career Development (CAREER) Award from the National Science Foundation. The five-year, $525,000 grant recognizes and supports promising young faculty for their exemplary research and teaching.
Research News asked Argall about her work’s challenges and potential.
Did robotics always fascinate you?
I didn't actually start working with robots until I matriculated as a doctoral student at Carnegie Mellon’s Robotics Institute. And I didn't study medical robots: my entire dissertation and postdoc explored teaching robots how to control their movements. So the focus was very much on robot autonomy and machine learning. I started working in the area of assistive and rehabilitation robots when I joined Northwestern and RIC five years ago.
What especially appeals to you about your research?
The most exciting aspect is leveraging robotics autonomy to advance human autonomy. Our goal is to make people with severe motor impairments more able and independent, more autonomous. We do this by making the assistive machine they use easier to control by introducing robotics autonomy. With this autonomy, the machines can partly control themselves and assume some of the control burden from the human.
Robotics autonomy doesn't have much representation within the field of rehabilitation robotics, but there clearly is a lot of low-hanging fruit. The key is to be judicious in selecting which problem to tackle, so that the effort is both technically feasible and also addressing user need in a way that is likely to be adopted by the human. That is the driving force behind all my lab’s current core research projects.
What’s the biggest challenge to making transformative advances with robots in rehabilitation?
One of the biggest hurdles is figuring out how to appropriately share control with the human user — and maybe even changing this control-sharing automatically, over time. This is important since we know that the human will be changing, either regaining ability through successful rehabilitation, or else losing ability due to a degenerative disease.
How does one integrate the “machine” with the “human” in this shared model?
In our system, a human is always in the control loop. We know from this and other assistive domains that humans largely prefer to retain control when able. So we have a scenario where the human and the autonomy are actually sharing control of the assistive robot. Getting this right is tricky, and crucial. Taking control away from humans when they want control makes them less able, which is entirely not the point. Conversely, giving control back when an individual doesn’t expect it can be dangerous. Our intuition is that successful control-sharing will be tuned and customized to each person because each person is unique in their physical abilities and preferences. So we build into our software knobs that can be tuned to customize the control-sharing.
What would most surprise laypeople about robotics?
How not capable real robots are. Science fiction and Hollywood do us a disservice here, and even we researchers do ourselves a disservice in posting cherry-picked videos that show the one minute of successful operation and don't mention the hour failed attempts. As a result, robots seem much more capable than they really are. The truth is that robots do exactly what we program them to do, and when things go awry it is not something inherently nefarious on the robot's part, but rather an oversight on the human programmer's part.
Did you ever consider a different career?
Not as an adult. I have always been drawn to research. When I was younger, field work had a big draw: being an archaeologist, or working with Doctors Without Borders. I never pursued this direction seriously. I did take all of the pre-med requirements at college, just in case I wanted to become a medical doctor. Upon finishing my undergraduate study, it was pretty clear that research was my path.