Dr. Betty H.C. Cheng, Professor in the Department of Computer Science and Engineering at MSU, is developing techniques to address the assurance and trustworthiness of autonomous vehicles.
While the future of autonomous vehicles is both promising and exciting, there are many obstacles to overcome for them to safely operate on the roadways.
“With the increasing movement towards autonomous driving and autonomous transportation, assurance is of paramount importance,” Cheng said. “Uncertainty poses the greatest challenge to the widespread and safe use of autonomous transportation.”
Cheng’s research looks at two broad categories of uncertainty and is attempting to address them for safer autonomous vehicle operation. The first category of uncertainty is environmental uncertainty, which includes obstacles such as road and weather conditions and human behavior. The second category is onboard uncertainty, such as unwanted feature interactions between different onboard systems.
“It is impossible to codify all the possible scenarios that a vehicle may encounter when driving,” said Cheng. “Our research is developing techniques to identify and mitigate unsafe behavior due to both types of uncertainty.”
Cheng is working with industrial collaborators and researchers to identify different types and sources of uncertainty. Her lab is using a combination of evolutionary search techniques, mathematical analysis, model-driven engineering, and machine learning to explore and mitigate negative impacts of adverse conditions to the vehicles. An increasingly important area of her focus is how to provide robustness and resiliency to autonomous systems even in the face of cybersecurity vulnerabilities.
“Mobility is an exciting area, but we also have a lot of work to do in order to ensure safe and reliable mobility,” Cheng said.