HORIZN was designed to detect anomalies and analyze hot spots: a large crowd gathers, vandalism has occurred and the system could flag unusual patterns and help surface possible response cues. The system could also be useful in responding to other emergencies such as severe storms or floods. It could access datasets from the National Oceanic and Atmospheric Administration.
“We could use our model to figure out if a specific area of the city is going to be flooded or if there’s going to be severe lightning storms involved,” says Pandey. “This would generate action cues and help expedite the response process.”
Pandey focused on creating AI-based extreme weather analysis and public safety components using massive real-world datasets, while Nikolaenko developed video-based crime detection components. “The project stands out because it combines technical depth, collaboration and real-world relevance into one system with a clear purpose,” adds Pandey.
Jun Bai, assistant professor in the UC Department of Computer Science, says their project HORIZN is unique because of its integrated approach.
“Many AI systems focus on only one area, such as crime analytics, surveillance or weather warning systems,” explains Bai, who served as an adviser for the project. “HORIZN brings these together into a single platform for broader situational awareness. That combination makes it more practical and impactful for real-world public safety use.”