Solving Physical Problems with Robotics

Professor Julius Dewald believes there are two distinct ways to approach robotics. The first is to build a targeted robotic system and then determine how its movement and application changes in different environments. The second approach — and one he supports — is to identify a problem first and then try to develop a robot to solve that problem.

"We use robotics as an exploratory tool to understand movement deficits following brain injury," said Dewald, who chairs the Department of Physical Therapy and Human Movement Sciences at Northwestern’s Feinberg School of Medicine.

Dewald and his Northwestern lab research team focus on understanding discoordination of an individual's upper limbs after a stroke or cerebral palsy diagnosis, and then develop novel neurotherapeutic training programs to enhance limb movement control years after a brain injury. He routinely collaborates with students in Northwestern Engineering's Master of Science in Robotics (MSR) program, particularly those in MSR Program Director and Professor Todd Murphey's Interactive & Emergent Autonomy Lab. 

"We're developing robotics for rehab purposes," said Dewald, who also is a professor in Biomedical Engineering, director of two NIH training grants for engineers, and co-director of the Shirley Ryan AbilityLab - Northwestern University C-STAR clinical core. "We create virtual environments for people to learn to move in, and robotics are a very important part for us to make that possible."

Murphey's lab collaborates with Dewald's and uses Dewald's motor control, biophysics and clinical expertise, and experimental systems as an environment for developing and studying human-robot interaction. Murphey's lab develops algorithms that enable a robot to physically assist a person while avoiding too much assistance that undermines therapy goals.

"Working with researchers in clinical settings provides an incredible opportunity to use robotics in high-impact scenarios," Murphey said. "That clinical setting also helps us focus on what sorts of technologies can realistically help people."

Robots in this environment are often used to support people while they walk or use their arms within a workspace. Oftentimes, this helps keep them safe while they develop or regain their motor skills.  

While robotics play a major role in this rehabilitation process, it's important to recognize they are still only part of the process. Robots are most effective in helping a patient recover when the underlying pathophysiological causes are understood. Researchers like Dewald and his collaborators combine the capabilities of robots, clinicians-scientists, and engineers when creating therapeutic plans.

That is why Dewald always wants to understand the challenge he's trying to address before coming up with a solution. There's value in building a robot for fun, but when you're trying to help someone regain a physical ability they've lost, you want to make sure you're doing it effectively and safely, Dewald said.

"There's often a disconnect between the engineer who can build a complex robot and a clinician scientist," Dewald said. "There are opportunities to use engineering in innovative ways, but when you develop something, try to understand the underlying problem you're trying to measure and treat first."