CMU

In an effort to combat increased traffic fatalities, Pittsburgh has collaborated with Dr. Ding Zhao and his Safe AI Lab at Carnegie Mellon University. Financed with a $1.3 million Department of Transportation grant, their goal is to revolutionize road safety within Pittsburgh by 2026 dramatically. This effort is part of an umbrella safety initiative underwritten by the Bipartisan Infrastructure Law, designed to reduce traffic fatalities and encourage safer roads within America.

Dr. Zhao’s team will employ state-of-the-art AI technology involving large language models (LLMs) when conducting an in-depth investigation of Pittsburgh’s streets. Large language models are especially adept at processing large data sets and can readily identify hazard points at dangerous intersections and other risk areas for drivers as well as pedestrians. Informed by this data, Zhao’s team will work with city officials closely to develop and install safer streets with a view toward decreasing accidents and saving lives.

The group plans to use simulations as a way of assessing how different safety adjustments would work citywide, testing them virtually before they are made in real life. This approach allows for extensive planning as well as optimizing the success of adjustments once they have been put into use. Pittsburgh Mayor Ed Gainey wholeheartedly supports the initiative, claiming collaborations with entities such as Carnegie Mellon are a must if meaningful progress is to occur on road safety. Further, Dr. Zhao and his team are using a next-generation testing platform, SafeBench, to confirm how these countermeasures would operate. The platform can factor everything from today’s current driving conditions all the way down into what will exist on a futuristic road, such as autonomous vehicles. It enables them to look ahead and call out obstacles before they happen, and make Pittsburgh’s roads safer for everyone, both now and into the future.

By
Vraj Parikh