The National Science Foundation has awarded Professor Guanpeng Li in the Department of Computer Science a total of $600,000 to develop a series of algorithms that could enhance the safety of autonomous vehicles.

The project, titled “Collaborative Research: MELIOREM:  An Integrated Evaluation Cyberinfrastructure towards Safe and Dependable Autonomous Driving Systems”, aims to establish a robust testing platform, MELIOREM, to evaluate and enhance the safety of autonomous vehicles (AVs). This innovative infrastructure will simulate a variety of everyday driving scenarios encountered by AVs on public roads, allowing a thorough assessment of their performance across diverse conditions. By leveraging large-scale computing resources, MELIOREM will help developers identify potential safety vulnerabilities and conduct comprehensive testing of AV systems. 

Traditional methods for AV testing face challenges due to the significant computational demands required to simulate real-world scenarios, which often lead to state space explosions and complicate vulnerability detection. AV software stacks add further complexity, making it both time-consuming and technically challenging to pinpoint and rectify errors. MELIOREM is set to not only advance the field of AV safety, but also serve as a valuable tool for academia and industry, supporting a broader professional community in enhancing transportation safety. 

In addition to AV safety, Professor Li's research interests span HPC fault tolerance, lossy data compression for HPC, machine learning dependability, and approximate computing.