Certain roadways in Florida are being transformed into testing sites for artificial intelligence aimed at discovering ways to prevent accidents. In an effort to enhance safety, Broward County has partnered with the University of Florida to implement AI-driven algorithms that monitor vehicles, cyclists, and pedestrians through cameras. As part of a pilot initiative, researchers are studying methods to prevent collisions before they occur, leveraging AI to identify potential solutions.
Broward is the latest community in Florida to adopt this innovative approach, joining previous pilot programs conducted in Gainesville and Seminole County.
In the past, researchers had to wait for a crash to occur and then analyze the data to determine its cause and identify possible safety measures, explained Min-Tang Li, spokesperson for Broward's Traffic Engineering Division. However, this new method "enables a more proactive approach – allowing us to identify potential collisions in advance and gain a clearer understanding of how they happen."
Currently, six intersections along Stirling Road in Broward are under evaluation to track how frequently near-miss incidents occur, as intersections are known hotspots for crashes that can result in injuries or fatalities, according to a research report.
Analysis
The AI technology analyzes video footage from traffic intersections to detect areas where "near misses" occur, providing valuable insights to prevent future accidents. The AI technology can help identify patterns, such as the time of day when near-accidents are most common and whether drivers face more challenges making left or right turns, explained University of Florida professor Sanjay Ranka. "The ultimate goal is to use these near-miss incidents to reduce actual crashes and, ultimately, fatalities," he said.
According to Min-Tang Li, the AI system analyzes the behavior of vehicles, pedestrians, and cyclists by tracking their locations, lanes, speeds, and acceleration patterns. Previously, relying on eyewitness accounts provided only a partial understanding of incidents. Li noted that such observations can be unreliable, as factors like the witness's position can influence their perception of events. "We couldn't see the full picture, and obtaining objective data was challenging," he added.
Ranka highlighted a key takeaway from a Gainesville study, which revealed that students near a high school needed more time to cross a particular intersection. However, extending the crossing time all day would have disrupted traffic flow. The solution was to adjust pedestrian crossing times only during peak hours in the afternoon, ensuring both safety and traffic efficiency.
In Broward, the program has already led to changes, such as installing new signs reminding drivers to stop for pedestrians. The final findings from the study are expected by June, with plans to expand the initiative to other areas in the county, officials said.
