by Simon Washington
In my over 30 years in the transportation profession, I have observed that many of us who have devoted our careers to traffic safety—whether that be in policy, business, or technology—are motivated by extremely personal experiences of tragedy and loss. There are a few such experiences for me, and probably the most profound occurred during a group bicycle ride about 12 years ago when a young woman was hit by a passing truck, essentially right beside me, and died instantly—I had just met her literally 10 minutes before she died. I will unfortunately live with this memory forever, and it is a constant reminder of why I am passionate about improving road safety globally.
Each fatality and serious injury represents a preventable tragedy, every thoughtful critical conflict analysis could save someone’s daughter, son, spouse, or parent. Ours is not abstract statistical work—it is about preserving human life and preventing the immeasurable grief that traffic fatalities inflict on families and communities.
My contribution to the important work of improving traffic safety lies in the realm of technology. When my colleague and personal friend Professor Tarek Sayed proved (in the early 2000’s) that critical conflicts using the power of AI and video analytics technology could replace crashes to save lives, I saw an opportunity to revolutionize what is possible to address the global road safety problem. The opportunity to harness technology to escape our reliance on reactive safety and deliver a proactive methodology was irresistible. The opportunity in fact was so significant that I decided to leave a successful academic career in 2019 and co-found Advanced Mobility Analytics Group (AMAG) to develop the technology for maximum real-world impact. AMAG was empowered further as a result of the acquisition by Transoft in late 2024.
Digital Technology Is Revolutionizing Everything In Traffic Safety
Transoft Solutions’ approach to traffic safety lies in the complex, careful integration of advanced video technology, artificial intelligence, machine learning, as well as statistics and econometrics. While we (along with the traffic engineering industry) have made tremendous strides in these computational areas, the sensor technologies embedded within our transport networks are legacy technologies with severe limitations. Legacy traffic signals and loop detectors, as examples, are decades-old technologies with an inability to perform dynamically in many cases and to detect vulnerable road users as examples. These and other legacy technologies are contributing to the profession’s inability to achieve Vision Zero objectives and eliminate fatalities and serious injuries from our transportation networks—creating a technological gap. Moreover, the global reliance on using crash data to make reactive decisions for improving network safety is fundamentally flawed—our industry’s reliance on deaths and serious injuries to prevent future deaths and serious injuries must be abandoned without hesitation at the first opportunity.
This technological gap has led to significant opportunities for advancement. Transoft’s video analytics platform VERALYTIX has been created to fill a critical technological gap preventing state, regional, and local governments from achieving their road safety goals. The technology allows engineers, operators, and planners to measure and assess every single road user’s movement at a location and convert them into actionable intelligence. Where traditional sensors cannot tell us how many pedestrians use a crosswalk, how many cyclists pass through an intersection, or whether they are following safe practices, modern video analytics provides continuous, objective, and accurate monitoring of all road users and their risk of harm using critical conflict analysis. And the technology also applies to emerging road user categories like e-scooters and high-speed e-bikes that travel at speeds up to 50 kilometers per hour—emerging road user categories with elevated risk of injury. The societal value and impact of technology that measures, quantifies, and enables governments to mitigate crash risks before a road user becomes a fatality or serious injury statistic is profound.
The use of critical conflicts for measuring risk is profound—road user data is gathered more quickly, is more reliable, and is more abundant, and therefore identifies reliable risk trends more quickly compared to crashes. Video analytics can be used for continuous road user monitoring—alerting operators and traffic engineers about important safety and operational events—or diagnostically for fixed monitoring periods (e.g. one week), facilitating deep dives into how a site is performing compared to others. Both continuous and fixed-duration monitoring offer a paradigm shift in how today’s traffic professionals manage, operate, and plan for safer and more efficient transportation networks.
From Reactive to Proactive: The Critical Conflict Revolution
Critical conflicts—often referred to as ‘near-misses’—requires a change in mindset, understanding, and thinking about network risk. It is akin to the mindset change the medical profession underwent with many diseases over the past decade—a shift from intervention to prevention. A shift from a focus on the deterministic diagnosis of heart disease to a focus on mitigating factors that reduce the risk of developing heart disease. Similarly, critical conflicts are probabilistically related to crashes. The repeated occurrence of critical conflicts increases the risk of escalation to a crash. A single critical conflict at a location represents a lower risk than many critical conflicts at the same location. This probabilistic view represents a shift from crashes that are deterministic. Importantly, what distinguishes a critical conflict from a ‘normal’ interaction on the network is the accumulation of two decades of peer-reviewed research that has established a time threshold of between 1.5 to 3 seconds (near to average perception/reaction time) under which road users on a collision course are at increased risk of a crash. This makes perfect sense—when road users headed toward collision have little time to react to avoid the collision, some road users will fail due to perception, reaction, distraction, fatigue, or impairment related deficiencies.
This proactive conflicts-focused approach enables us to intervene before tragedies occur, fundamentally altering our relationship with traffic safety from reactive response to preventive action. This approach aligns with the Safe System Approach being adopted globally.
The implications extend far beyond individual intersections and corridors. When we can identify and quantify risk patterns across entire networks, we can prioritize safety investments more effectively, allocate resources where they will have the greatest impact, and demonstrate measurable improvements in safety outcomes without waiting years for crash data to accumulate.
Integrating Video Analytics & Design
What sets Transoft Solutions apart in this rapidly evolving field is our company’s foundation in traffic engineering expertise rather than pure technology development. While numerous companies have applied AI and machine learning expertise problems to transportation, Transoft is comprised of traffic engineers and transportation safety professionals who understand the industry problems intimately and have leverage convenient technology to solve them.
This ground-up perspective, built on more than 200 years of collective thought leadership in traffic safety, positions Transoft uniquely to understand how customers need information presented and what insights they require to make sound safety decisions. Our team doesn’t just develop technology; we understand the practical challenges that transportation professionals face daily and design solutions that address real-world needs.
The integration of our video analytics capabilities with groundbreaking design tools creates unprecedented opportunities for comprehensive safety solutions. Through our recent acquisitions, including Advanced Mobility Analytics Group, Transoft can now offer clients network screening capabilities, before-and-after safety evaluations, measurement of risk profiles before improvements and quantifying safety benefits after implementation. These capabilities allow agencies to demonstrate the effectiveness of their investments with a detailed evidence base to build business cases that will deliver outcomes for their stakeholders.
Looking ahead, Transoft is developing even more sophisticated integrations between our analytics and design platforms. Our vision is to use video analytics to identify specific risk patterns at a location, then leverage our design tools like TORUS Roundabouts to create targeted improvements that directly address the measured safety concerns. This closed-loop approach—from risk identification through design optimization to post-implementation validation—represents a level of comprehensive safety management that was previously impossible.
Trust Requires Effort
Video analytics data sounds fantastic…but can it be trusted?
All sensors and technologies have limitations. Inductive loops, LiDAR, radar, and video all have limitations. Machine learning and AI also have limitations. The question we all need to ask is whether the technology is accurate enough to answer the questions the industry is seeking answers to. In this context, Transoft is committed to building, maintaining, and developing the most reliable and accurate analytics platform for addressing our customers’ specific transportation safety, operations, and planning needs. To this end, we have developed a rigorous, multi-layered process that combines technological innovation, independent validation, and a commitment to transparency with our customers. We support an independent Research & Development team within Transoft to maintain our leading position in the industry. We remain engaged with research institutions and leading researchers throughout the world to test, validate, and refine our technologies. Finally, we are actively engaged in professional societies, conferences, and associations to engage and exchange ideas and feedback with our customers and stakeholders.
Our technical validation process is unparalleled. It begins with precise calibration of video systems, ensuring that each camera’s field of view is accurately geolocated. This foundational step enables the system to reliably track every road user across complex intersections and corridors. To verify accuracy of our AI, we systematically validate road user counts and classifications against manual observations using 15-minute intervals of human-coded data—compared to automated results. This validation is used to either refine our AI or validate and quantify for our customers its performance in the field. Our patented error-filtering technology automatically identifies and removes common sources of error in video data, such as occlusions, camera vibrations and movements, lighting changes, or sensor noise for improved downstream analytics accuracy and reliability. We have implemented thermal camera imaging for the most challenging night-time conditions. These innovations collectively ensure that the analytics platform delivers high-confidence results, even in the most challenging real-world environments.
Beyond technology, Transoft emphasizes openness and customer education to encourage trust. We’re transparent about the strengths and limitations of video analytics, and we work closely with clients to interpret results and apply them effectively. This collaborative approach helps agencies and consultants build trust in the data, supporting a cultural shift toward proactive, data-driven safety management.
Through this combination of calibration, validation, patented error filtering, and customer partnership, Transoft ensures its video analytics are not only technologically advanced but also trusted and actionable for safety-driven design.
The Path to Vision Zero
The ultimate goal that drives our product development in Transoft is Vision Zero, i.e., the elimination of traffic fatalities and serious injuries—the metric that matters most. This is not merely an aspirational target but the most sensible goal of a civilized society. Importantly, we will never get to zero deaths without proactive technology. Proactive technology is not just helpful for reaching Vision Zero, it is essential.
The transition to automated driving will create additional complexities that reinforce the importance of comprehensive safety monitoring. During the inevitable mixed-traffic period, when automated and human-driven vehicles share the same roadways, we as a profession will need to understand how these different driving behaviors interact and what new risk patterns emerge. Video analytics technology will be crucial for continually monitoring and managing this transition safely.
At Transoft Solutions, we are committed to developing and deploying the technology, expertise, and partnerships necessary to make Vision Zero achievable. The tools exist, the knowledge is available, and our commitment is unwavering.
—
Simon Washington is a transportation safety expert and co-founder of Advanced Mobility Analytics Group, which was acquired by Transoft Solutions in 2024. With 25 years of academic research experience at leading institutions including UC Berkeley and Georgia Tech, he now leads Transoft’s efforts to transform traffic safety through proactive video analytics technology.