News, field results and a few honest opinions.

A 24-hour Georgia DOT field evaluation put Currux Vision through the same four-part test the state uses to qualify any video detection system — missed calls, false positives, dropped calls, and count accuracy — using a single camera to cover an entire intersection rather than one sensor per lane. The results: 99.95% call detection accuracy, zero dropped calls in 6,491 actuations, and 96.38% count accuracy against manually verified ground truth. GDOT has since approved Currux Vision for video-based vehicle detection statewide.

In rain, fog, night and glare — the exact conditions that used to justify stacking on radar or LiDAR — Currux Vision's AI, reading the same video feeds agencies already have, is now validated at 99.6% detection accuracy.

Within one week of activation across Pinellas County, Florida, the AI-based Corridor Control platform cut traffic delay and corridor emissions by more than 25%.

LiDAR and radar are modern, but they're still blind to context. Here's why validating AI against the actual image, not a proxy for it, changes what a detection system can do.

A clear-eyed look from The Economist's Tim Cross at where machine learning genuinely delivers, and where the industry's hype has outrun it — shared because it's a useful gut check.

InnovationMap's Natalie Harms covers the SmartCity ITS testing results out of San José, California, and what a 99% accuracy figure meant for an early-stage Houston company.

Extensive testing with the San José Department of Transportation confirmed up to 99% accuracy for SmartCity ITS, across day, night, rain, camera vibration and partial obstruction.

Independent testing of the joint Costar Technologies and Currux Vision traffic video analytics platform confirmed 95–98%-plus accuracy across day, night, rain and partial obstruction.

Currux Vision partnered with camera manufacturer Costar Technologies to bring an integrated AI traffic video analytics platform to DOTs, municipalities and private infrastructure operators.

InnovationMap's Alice Levitt tells the origin story behind the Currux name — and how founder Alex Colosivschi went from energy finance to building a smartphone driver-coaching app.

How a phone mounted on the dash can deliver the same collision and lane-departure warnings as dedicated ADAS hardware — at a fraction of the cost, for every vehicle in a fleet.

Telematics laid the foundation for fleet safety. AI, computer vision and mobile apps are what's letting fleet managers act on it — without a hardware budget.

A behind-the-scenes conversation on what it took to run three machine learning models at once, in real time, on a phone bolted to a windshield.