In an effort to improve how those systems operate, Kennesaw State University computer science researcher Kun Suo earned a $356,782 National Science Foundation (NSF) grant focused on developing more efficient and secure computing infrastructure for intelligent vehicles.
The project, “CSR: Co-Designing Virtualization and System Infrastructure for Efficient Vehicle Computing and Safe Onboard Intelligence,” explores how vehicle hardware and software can work together to process complex information while reducing delays and protecting critical driving functions.
Unlike traditional computing environments, however, a delay or failure inside a moving vehicle can have immediate consequences.
Efficient and Fast
“The new vehicles today are more and more like a computer,” Suo said. “We want to make the vehicle computing very, very efficient, very, very fast, and use fewer resources.”
Suo, an associate professor of computer science in Kennesaw State’s College of Computing and Software Engineering, is examining how those competing demands can be better managed.
A key component of the research is virtualization, which allows multiple software applications to share the same physical computing hardware while operating within separate environments. Suo’s project seeks to create frameworks that more effectively separate safety-critical vehicle functions from less critical digital services.
The goal is to help the vehicle decide how its computing power should end up used at any given moment. Systems tied directly to driving and safety would be given greater protection and priority, while less critical features could continue operating without interfering with those essential functions.
Kun Suo wants to develop more efficient and secure computing infrastructure for intelligent vehicles.
Source: Kennesaw State University
The research also addresses how those systems can operate securely.
By creating stronger separation between critical driving functions and secondary applications, Suo and his team want to reduce the chances that problems in one area of a vehicle’s computing system could interfere with another.
“The biggest challenge here is to know the hardware and know the software and know how the entire system stacks so that we can make it more efficient and safer,” Suo said. “I can bring my older knowledge and experience into this new topic.”
The researchers are approaching the challenge from the hardware and software sides.
Improving AI Efficiency
One part of the project focuses on improving the efficiency of artificial intelligence workloads, including how computing tasks end up scheduled and how information moves through the system. Another explores security mechanisms that can better isolate applications with different levels of importance.
Suo’s work builds on his previous research examining edge computing, which processes information closer to where it ends up generated rather than sending it to a distant cloud server.
Suo sees the project as part of a larger shift in how computing will shape transportation.
“I’m very excited about vehicle computing and onboard intelligence,” Suo said. “I believe there will be more intelligence in the future and also more computing integrated into the vehicle platform.”