Enhancing Urban Road Intersection Safety

🚲 Mobility

📍 New York City, USA

About

Status:

✅ Finalized

New York City is the largest city in the United States and a global hub for finance, culture, media, and innovation. Comprising five boroughs—Manhattan, Brooklyn, Queens, the Bronx, and Staten Island—it is home to over 8 million residents. NYC plays a central role in the global economy, hosting major institutions like Wall Street and the United Nations. The city is also a leader in smart urban services, sustainability initiatives, and resilient infrastructure.

🗓️ Start: 2024-03-30

⇨

🗓️ End: 2025-03-31

Urban intersections are high-risk zones where diverse users like cars, pedestrians, cyclists, and micro-mobility, interact in complex ways. This challenge seeks a low-latency, privacy-compliant solution using edge computing and AI to detect and analyze dangerous situations in real time. By processing data locally and sharing only anonymized insights, the system will improve safety, enable rapid responses, and support data-driven urban planning.

Description of the Challenge

In Manhattan, intersections concentrate highly complex interactions between different types of road users. This density and complexity significantly increase the risk of accidents and dangerous situations. When accidents do occur at these crossroads, many result in severe injuries or fatalities. The problem affects all road users, but particularly vulnerable groups like pedestrians and cyclists, and creates major operational and planning challenges for the municipality.

Current situation and gap

Improving safety at crossroads is a major priority for municipalities. However, identifying risk situations and understanding the dynamic of traffic interactions at busy intersections remains highly challenging without advanced monitoring tools.

Currently, relying exclusively on distant cloud infrastructure to process real-time video analysis is insufficient due to high latency. Furthermore, transmitting raw video data from public spaces to analyze traffic raises significant ethical and privacy concerns. The city lacks a privacy-compliant, low-latency method to actively monitor these intersections, analyze the data in real-time, and use it for long-term infrastructure planning.

Expected Outcomes

New York City needs the capability to monitor and analyze complex traffic interactions at urban intersections in real-time to detect potentially dangerous situations. The city needs to achieve this with very low latency while strictly protecting citizen privacy, meaning they cannot rely on transmitting raw video footage to distant cloud servers.

Expected outcome

Desired solution: An integrated system combining video detection, edge computing infrastructure, and artificial intelligence (AI) algorithms.
Functional expectations: The solution must process video streams locally at the intersection level (edge computing) to detect risk situations and analyze traffic behavior. It must ensure that only anonymized data is transmitted to a central digital environment or hypervisor (such as the Kentyou Eye platform) for further processing. The system should integrate multiple data sources (video, sound recognition, IoT devices) and support both real-time monitoring dashboards and historical data analysis.
Expected impact: Enhanced situational awareness and improved safety at urban intersections. By providing a clear understanding of traffic behavior and risk situations, the solution will support immediate operational responses and enable data-driven decision-making for long-term urban mobility management and infrastructure improvements.

Space for Solutions and Experimentation

Available experimentation space: Urban road intersections in Manhattan, New York City, USA.
Scale of experimentation: Intersection / Neighbourhood scale.
Available data, digital tools, technical support, city engagement: The experimentation is supported by a strong consortium of partners including Kentyou, Columbia University, Rutgers University, New York City, Santander, and Universidad de Cantabria. Access is available to the Kentyou Eye platform (smart city hypervisor) to visualize traffic conditions and analyze historical patterns.
Any specific conditions or constraints for the experimentation: Strict privacy and technical constraints apply. The system must operate with very low latency (necessitating edge-based data processing) and must ensure the protection of privacy by not transmitting raw video footage, but rather localizing the analysis and sending only anonymized data.
Existing experience with pilots or tests: The project has successfully demonstrated the feasibility of combining video detection, edge computing, and AI to analyze traffic interactions in real-time while respecting privacy constraints.