USDOT Taps Derq for AI-Powered Intersection Safety Systems

DETROIT – October 22, 2026 – The U.S. Department of Transportation (USDOT) is accelerating a fundamental shift in how the nation manages its urban roadways, moving away from post-crash data analysis toward real-time, predictive intervention. In a significant endorsement of commercial artificial intelligence in public infrastructure, the USDOT has awarded Detroit-based intelligent transportation company Derq a federal contract under Phase 1 of its Intersection Safety Systems (ISS) initiative.

Following a highly competitive, multi-year evaluation process, Derq was selected as one of only three teams to demonstrate advanced systems designed to identify, predict, and mitigate dangerous situations involving vehicles and pedestrians before a collision occurs. The award draws from a $5.5 million Phase 1 funding pool dedicated to prototype development and closed-track testing, positioning the company at the forefront of a federal mandate to eliminate roadway fatalities.

For institutional investors and municipal bond analysts monitoring the smart city sector, the contract represents a critical maturation point. The federal government is officially signaling that predictive edge computing and vehicle-to-everything (V2X) communication are no longer experimental concepts, but necessary components of federally funded transportation infrastructure.

From Post-Crash Data to Active Intervention

Historically, traffic safety enforcement has relied heavily on reactive measures: red-light cameras that issue tickets after a violation, or crash data logging that helps engineers redesign intersections years after a cluster of fatalities. The USDOT's ISS initiative, which builds upon the rigorous 2023-2025 Intersection Safety Challenge, demands a proactive approach.

Derq stands out in this federal push as the only private corporate lead selected among the three Phase 1 awardees, competing alongside and collaborating with traditional academic and public sector consortia. The company previously earned a Tier 1 placement in Stage 1B of the challenge, securing $750,000 after successfully demonstrating its system's high-precision detection and conflict prediction capabilities using USDOT-provided sensor data.

“Most systems today tell you what happened. Ours also predicts what’s about to happen and acts on it,” said Dr. Georges Aoude, co-founder and CEO of Derq. “As the only private company selected for this USDOT program, with technology already running at intersections across the country, Derq is ready to prove that real-time crash prevention can become part of everyday traffic operations and protect communities at scale.”

Unlike traditional systems, Derq’s platform fuses real-time camera footage with radar, traffic signal phase and timing (SPaT) information, and V2X messages. By processing these inputs locally at the roadside, the AI anticipates imminent conflicts—such as a distracted driver accelerating toward a red light while a pedestrian steps into the crosswalk. The system can then instantly trigger a roadside warning sign, broadcast an alert to connected vehicles, or even preemptively adjust the traffic signal to hold all lights red until the hazard clears.

Combating the Pedestrian Safety Crisis

The urgency behind the USDOT's investment is underscored by grim national statistics. Traffic crashes remain a severe public health crisis, accounting for an estimated 36,640 U.S. fatalities in 2025. Intersections are particularly lethal, representing approximately one-quarter of all traffic fatalities and roughly half of all traffic injuries annually.

Vulnerable road users bear a disproportionate share of this risk. Intersections account for about 18% of all pedestrian fatalities nationwide. With pedestrian deaths having recently reached their highest levels since the early 1980s, federal regulators are increasingly desperate for scalable technological interventions that align with the National Roadway Safety Strategy's vision of zero roadway deaths.

To ensure the technology can perform under the unpredictable conditions of urban traffic, Phase 1 testing will be conducted in highly controlled but realistic environments. Derq will deploy its prototypes at the Virginia Tech Transportation Institute’s Virginia Smart Roads and the Texas A&M Transportation Institute’s RELLIS Proving Grounds. These access-controlled testbeds allow engineers to simulate complex, high-risk scenarios—such as low-visibility conditions and erratic driver behavior—without endangering the public.

Sub-Second Latency at the Edge

The technological hurdles required to alter traffic signal states before a crash occurs are immense. The primary challenge is latency. To effectively warn a motorist or change a traffic light in response to a vehicle traveling at 45 miles per hour, the system must detect, classify, predict, and act within fractions of a second.

Industry analysts note that relying on centralized cloud computing is functionally impossible for this application; the round-trip time for data to travel from an intersection camera to a remote server and back is simply too slow to prevent a high-speed collision. Consequently, the USDOT program heavily favors edge computing architectures.

Derq addresses this by processing the fused sensor data directly at the roadside cabinet. The system is built on a robust portfolio of 23 patents spanning sensor fusion, road-user intent prediction, V2X-based warnings, and real-time conflict mitigation. Independent edge compute engineers emphasize that the ability to accurately predict trajectory and intent—differentiating between a vehicle that is braking hard for a yellow light versus one that is accelerating to run it—is the defining metric for success in the ISS program. False positives that unnecessarily disrupt traffic flow are just as detrimental to municipal adoption as missed detections.

Market Opportunity and Strategic Partnerships

Founded in 2016 as a Massachusetts Institute of Technology (MIT) spinoff, Derq has rapidly transitioned from an academic research project into a primary federal contractor. This evolution highlights a lucrative market opportunity for tech firms capable of bridging the gap between advanced AI research and legacy municipal infrastructure.

The public-private syndicate assembled for this deployment illustrates the massive scale of the smart mobility sector. Derq's public-sector partners for the upcoming deployments include the Texas Department of Transportation (TxDOT), the Virginia Department of Transportation (VDOT), the City of Houston, and the Virginia cities of Alexandria and Falls Church. These agencies are not merely passive observers; entities like TxDOT have been actively involved in earlier stages of the USDOT challenge, signaling a strong institutional appetite for procuring and scaling these systems once validated.

Crucially, Derq has also aligned with major corporate players to facilitate deployment. Industry partners include global engineering firm WSP USA, the Intelligent Transportation Society of America, and telecommunications giant AT&T. The involvement of AT&T is particularly notable, following the telecom's strategic investment in Derq and an agreement to integrate Derq's intelligence into AT&T's Intelligent Transportation Platform. This partnership provides the essential low-latency connectivity backbone required for V2X communications across wide geographic areas.

Looking ahead, the financial implications of the ISS program extend well beyond the initial $5.5 million Phase 1 allocation. The USDOT anticipates releasing an additional $11.5 million for Phase 2, which will transition the successful prototypes from closed test tracks to real-world intersection deployments. Initially operating in an observation-only mode, these Phase 2 deployments will eventually introduce live mitigation actions, provided the safety integrity of the AI can be conclusively proven to federal contracting officers and local transit authorities.