Shima Nazari Drives Innovation in Mobility with Expertise in Control Systems

With the advent of autonomous driving and constantly changing safety standards and emissions regulations, the vehicular world is quickly driving toward a brave new world. Shima Nazari, an assistant professor of mechanical and aerospace engineering at the University of California, Davis, who is steering projects like the electrification of big machinery and advanced decision-making for driverless cars, aims to map this uncharted territory.

From Giant Batteries to Game Theory

Working with the California Department of Transportation, or Caltrans, Nazari and her team are investigating how to electrify the department's fleets of work vehicles, like loaders, excavators and street sweepers.

These types of off-road vehicles have different needs than the light-duty, on-road vehicles most people drive. They are much larger, with huge, complex systems, and how they are used is drastically different from an everyday car.

"With a car, you drive it around an hour per day, but a loader is supposed to work eight hours per day and would need a fast-charging cycle," said Nazari. "So, how would the battery life be different in this system, how can we develop control systems that are aware of battery aging? These are the problems we are working on."

Those batteries also have to be charged somewhere. Cars can stop at charging stations out on the road, but on a building or construction site, heavy-duty machinery can't get to those sources. Nazari's team is looking into mobile chargers that can be moved from site to site; these table-sized batteries would be able to hold up to 500 kilowatt-hours of energy. (An average electric car has a battery size of 40 kilowatt-hours.)

Elsewhere, Nazari's team is using data-driven methods and game theory to find efficient decision-making solutions for when an autonomous vehicle, or AV, has to interact with a human driver, which, mathematically, is a difficult problem to solve.

"When you are making a decision in a static environment where everything is stationary, you simply need to see where the obstacle is versus your possible paths and design the best one," said Nazari. "But when you are dealing with another intelligent agent in an environment where you have conflicting goals, every action you take will cause a reaction from them. You have to find the best actions to optimize your objectives, regardless of the other driver's free action."

Current methods for this type of decision-making depend on heavy computations and are useful for simple systems with a few agents. With the complex nonlinear dynamics of AVs, however, the problem becomes more complicated, and the existing solution methods are no longer useful. 

Going Off the Beaten Path

From a young age, Nazari showed an interest in machines. Her parents, she says, were keen on her going into medicine, but she was more interested in how aircraft and cars work.

"I grew up in the Middle East in a traditional family," she said. "Before me, there were no female engineers, so I think they didn't know what the role of a female engineer would be or what type of job would be appropriate for a girl."

Nazari pursued an emphasis in math in high school and attended the Sharif University of Technology for her bachelor's and master's degrees in mechanical engineering, focusing on dynamic modeling and energy conversion in jet engines to be used in control system development.

At the University of Michigan, she earned a second master's degree in electrical engineering and a Ph.D. in mechanical engineering. Her research focused on powertrain control and energy management for hybrid electric vehicles, showing that while human-driven hybrid electric cars require large motors to recover the energy from aggressive braking, future autonomous hybrid vehicles could benefit from small electric motors because of their extended reaction times.

Working the System

Specializing in control systems today is quite a bit more involved today than it would have been 50 years ago. Back then, cars were simpler. Now, standard functions like power windows and airbag deployment, as well as features like automatic emergency braking, powertrain and adaptive cruise control, require thousands of lines of code.

"The lines of code that are in the different electronic control units of [a typical] car are more than a jet fighter," said Nazari. "In a high-end luxury car, the lines of code are larger than the Facebook platform."

For Nazari, though, finding solutions to these problems in vehicle control systems and finding ways for vehicles to reach their full potential — from finding a path to electrifying big machinery to better decision-making for safer and more efficient AVs — is the most exciting part of the job.

"Control systems are really my thing," she said. "It has all the components of mechanical engineering — thinking about how different parts of the system have to work together to get the best outcome, the best efficiency, the best performance — and it is all about math, which I love the most."