NYC DoT expands use of AI activity sensors to 100 locations

New York City Department of Transportation (NYC DoT) has deployed new cutting-edge artificial intelligence (AI) sensors to better understand how New Yorkers use their streets and to inform safer street design.

First piloted in 2023, Vivacity’s small, privacy-protective smart traffic monitoring sensors will count pedestrians, cyclists, buses, and vehicles, and assess patterns in how people use the streets. Sensors will be installed at approximately 80 additional locations across the five boroughs to accelerate data-driven street safety improvements.

Building on the successful pilot, NYC DoT will this year scale up a technology that has already transformed how the agency measures traffic activity. Initially installed at 20 locations, the sensors replaced traditional manual traffic counts with continuous, real-time data collection – providing a far more comprehensive picture of how pedestrians, cyclists, and vehicles move through city streets.

“Safer street design starts with understanding what is actually happening on the street,” said NYC DoT commissioner Mike Flynn. “These high-tech sensors will help us evaluate how people are walking, biking, and driving so we can design safer streets and encourage safer behaviour.”

Mounted on NYC DoT’s street infrastructure, the sensors anonymously analyse street activity. In addition to counting different types of road users, sensors can measure speeds, capture turning movements, and map how different users move. For example, the sensors could identify areas where pedestrians are crossing mid-block instead of at crosswalks. This could help identify locations where a mid-block crosswalk would be beneficial.

The agency reports that traditional planning methods have previously relied on short-term, labour-intensive counts. The sensors operate continuously, capturing changes in travel patterns by time of day, season, and street design.

Designed with privacy as a core value, video footage is processed in real time and immediately discarded, with only anonymous data retained. Faces and license plates are deliberately obscured in any video footage.