Case Study: Near-Miss Collision Analysis

University of California, Los Angeles

In January 2025, Sensagrae extended its license agreement with UCLA to provide our computer vision AI camera software to provide real-time traffic and safety analytics. Our SensaVision Edge solution, which connects to 4 Axis Communication cameras and a NVIDIA Jetson device, provides real-time data and analytics. UCLA’s Transportation Department uses the near-miss collision data and analysis to gain insights into engagements between vehicle-to-vehicle, vehicle-to-non-vehicle, and non-vehicle-to-non-vehicle events. 

Key Achievements:

• Provide proactive insights to safety events from near-miss collisions and wrong-way driving.
• Provide a digital twin simulation to recreate the events of the near-miss for the customer to review. 
• Provided a feedback loop to allow the customer to verify near-miss collision events and allow a human-centric element for the customer to approve events and provide feedback on their expectations of how they classify near-misses from their perspective. 
• Actionable Insights: Automated process for near-miss collision events and recommendations from, USDOT’s Federal Highway Administration (FHWA), Highway Safety Programs, Proven Safety Countermeasures initiative (PSCi) for signalized intersections. This data enables customers to review countermeasure recommendations, identifying solutions to implement and reduce near-miss collision events. 

Objectives/Project/Challenge:

• In 2028, UCLA will host the Olympic Games and host the Olympic Village. With the Olympics’ goal to make the venues pedestrian-friendly and safe, UCLA is assessing traffic counts and volume, as well as vulnerable road user behavior, movements, and safety, to help improve traffic flow and drive mobility improvements at one of its busiest intersections on campus. 
• Provide 24/7 365 days of continuous traffic counts, turning movements, and safety events. 
• Provide actionable insights that inform customers with recommendations, saving them time and money by eliminating the need to assess data and determine how to prevent safety events. 
• Leverage the traffic and turning movement counts to improve signal phase and timing performance and scheduling. 
• The emerging technology allows UCLA Transportation to evaluate the benefits of emerging Intelligent Transportation System technology.  

The Solution:
Sensagrate worked with UCLA Transportation and met monthly to discuss the data, obtain feedback, and share new product features. The UCLA Facility Department deployed the technology, and Sensagrate’s Field Engineer assisted with a successful installation. 

Results:

• UCLA can identify peak-miss times using traffic volume counts, noting periods from 6 AM to 10 AM and from 4 to 6 PM. 
• UCLA was able to identity eight countermeasures to implement improvements to reduce and prevent near-miss collisions. The safety type events of countermeasures include: Geometric Modification, Lighting, Signal Modification, and Ped and Bike Safety Improvements. 
• In three months, the total number of events was detected:
  • 311 Post Encroachment Time (PET) with 0.3 as the Highest PET Score Rating
  • 61 Time To Collision (TTC) with 0.1 as the Highest TTC Score Rating
• Tracking maps of events with heatmaps provide insights into the locations at the intersection with the highest risk volume. ​