Traffic crashes remain a leading cause of preventable deaths globally. In the United States and Canada alone, 42,000 people lose their lives in traffic crashes every single year1. Tragically, vulnerable road users—pedestrians and cyclists—bear a disproportionate burden, accounting for 26% of all traffic fatalities worldwide2.
For decades, transportation professionals have worked tirelessly to reverse this trend, yet the traditional methods of evaluating road safety are structurally flawed. It’s time for the industry to move away from reactive safety models and embrace a proactive approach.
At the core of this shift is Vision Zero, a global initiative built on the premise that no loss of life on our roadways is acceptable. Rather than viewing traffic deaths as inevitable bi-products of mobility, Vision Zero recognizes that humans naturally make mistakes and shifts the burden of responsibility to system designers and policymakers. For engineers, this means building a “forgiving” infrastructure where human error does not result in fatal or severe consequences. However, actually achieving this systemic paradigm shift requires a fundamental evolution in how we diagnose danger on our streets.
The Fatal Flaw of Historical Crash Data
Conventionally, traffic engineers have relied heavily on historical crash data to pinpoint dangerous intersections and justify safety funding. However, this method has critical limitations that prevent cities from getting ahead of the curve:
- It is inherently reactive: Relying on crash history means we are fundamentally waiting for people to suffer harm or lose their lives before taking corrective action.
- It is slow: It can take between five and ten years for statistically significant crash trends to appear in the data. By the time engineers recognize a pattern, the operational reality of the intersection may have entirely changed.
- It lacks context: Crash reports frequently miss key pieces of behavioral context—or go entirely unreported—making it impossible for safety engineers to determine the true root causes of conflicts.
A spreadsheet simply cannot illustrate how an eastbound driver, hyper-focused on shooting a gap in oncoming traffic, narrowly executed a “left hook” turn right into the path of a pedestrian.
A Proactive Shift: Miovision’s AI-Powered Near-Miss Solution
To truly make a dent in traffic fatalities and achieve Vision Zero goals, we need diagnostics that tell us where crashes will happen, not just where they have happened. Enter safety studies.
Miovision Safety Studies combine proven surrogate safety methodologies with cutting-edge artificial intelligence to evaluate risk before collisions occur.
When video data is processed, our computer vision AI detects and classifies various road users, utilizing tracking algorithms to link their movement frame by frame. The software translates the video’s pixel space into real-world coordinates, creating highly accurate spatial mapping. By analyzing the intersecting trajectories of road users, the AI evaluates interactions based on critical risk factors such as vehicle speed, impact angle, user type, and temporal separation.
The predictive power of this technology is staggering. Miovision’s near-miss data can predict with 94% accuracy whether an injury collision will occur at a specific site within the next five years.
Measuring Success in Weeks, Not Years
Armed with near-miss video analytics, agencies can move from generalized safety updates to highly targeted, precision countermeasures. Because near-miss data delivers deep insights into specific conflict configurations, cities can implement high-ROI solutions without unnecessarily disrupting traffic flow where it is already safe.
Crucially, safety studies allow engineers to measure the effectiveness of safety treatments almost immediately. Instead of waiting another ten years for crash data to validate an infrastructure update, cities can implement a change, deploy cameras for a few days, and receive a comprehensive safety study back from Miovision within 14 days.
The real-world results speak for themselves. In a recent project spanning 50 intersections across five Canadian cities, engineers used Miovision data to implement 119 low-cost countermeasures—such as pavement marking updates and signal adjustments. This targeted approach yielded an incredible 85% average reduction in risk for the studied conflicts. In Vancouver specifically, these proactive adjustments led to a 55% reduction in near-miss events involving cyclists, contributing to a 68% surge in cycling volume.
When we make streets objectively safer, we invite the community to utilize active, multimodal transportation with confidence. By prioritizing proactive near-miss data, transportation professionals finally have the compelling visual evidence and rapid feedback loops required to save lives and make Vision Zero a reality.
1NHTSA Press Release 2022
2WHO, Global Status Report on Road Safety, 2023
