By: Brandon Pytel
Date: April 9, 2018
If Cars Only Had a Brain: CEAS PhD Student Receives Award for Bio-inspired Robotics Research
By: Brandon Pytel
Mechanical engineering PhD student Mohammad Sarim receives the GSGA Graduate Student Award for his work in bio-inspired navigation and localization in robots.
Whether you’re conscious of it or not, when you go into a room, your brain records its layout — the walls, the floor, the furniture — and prepares your body to navigate through it.
As humans, we learn from our environment and make rational decisions based on that knowledge. For robots, and self-driving cars, it’s not that easy.
The human brain has billions of neurons that makes entering a room — or riding a bike or driving a car — something we don’t even think about. But robots don’t have a human brain. Researchers at the University of Cincinnati (UC) developing robots to do the tasks of humans have to overcome this big problem.
Mohammad Sarim, a PhD student in UC’s mechanical engineering program, has spent the last several years developing brain-inspired navigation and localization solutions for robots. He has worked under the guidance of his adviser Manish Kumar, PhD.
Sarim’s work recently earned him the Graduate Student Governance Association’s Graduate Student Award for Exemplary Scholarship in Physical Sciences and Engineering. The award, which includes a monetary prize, recognizes one student each year who exhibits the highest degree of scholarship in his or her respective field of study.
“It’s a great compliment to receive this award,” said Sarim. “As a researcher, this honor encourages me to continue seeking knowledge.”
Sarim simulates memory devices and integrates these devices into robots. In other words, he replicates brain behavior in robots. He has shown that robots coupled with memory devices can learn and memorize their environment and respond accordingly.
“It’s like a baby learning how to walk,” says Sarim.
Because of their limited bandwidth, conventional computers cannot easily replicate synaptic behavior of the complicated neural network of the brain. The synaptic memory devices integrated into a ground robot, however, can sidestep this problem while using considerably less energy.
Sarim’s research can be applied to other industries that rely on artificial intelligence, like those industries developing self-driving cars.
“Self-driving cars must adapt to their environment, making decisions about path planning while avoiding obstacles,” says Sarim. “The current self-driving cars can perform very well in pre-programmed scenarios. When there are thousands of such cars driving around, however, these vehicles need to make decisions in split-second situations that they may not have been trained for.”
Sarim’s research can help these cars make those decisions.
Over the last four years, Sarim has enjoyed developing his research while taking advantage of UC’s state-of the-art facilities.
“UC has some of the best resources for robotics research,” he says. “The College of Engineering and Applied Science’s excellent lab facilities, outstanding faculty members and welcoming environment for international students make UC one of the best universities in the country.”
After graduation, Sarim, who successfully defended his dissertation earlier this month, is moving to Troy, Michigan, to work full-time with Aptiv’s self-driving car unit. Says Sarim, “I hope to apply my knowledge to progress robotics research that benefits humanity.”
With more self-driving cars hitting the road every year, Sarim’s research at UC will help create a safer, and smarter, commute.