Edge AI can detect road issues, but without precise positioning, the data falls apart. See how Dareesoft and Swift Navigation solved this in a real city deployment.
Watch our joint panel on AI-driven road asset monitoring
Recorded at NJ TransAction 2026, where transportation agencies and industry leaders share real-world deployments.
Swift Navigation and Dareesoft shared a real deployment with a major U.S. city, using edge AI and precise GPS to monitor speed limit signs at scale.
The system was built around Dareesoft’s Road Analyzer, a dash-mounted camera that detects and reports road assets as vehicles move through the network.
The goal was simple: identify missing, damaged, or obscured signs and keep the inventory up to date.
The Problem
The AI worked. It could detect signs and flag issues. But the system could not place them reliably.
With standard GPS, each pass recorded a different position for the same sign. Over time, a single asset could appear as many. That made it difficult to:
- Deduplicate detections
- Track changes over time
- Reconcile against existing databases
More fundamentally, each drive was captured in isolation. Every pass produced a slightly different version of the same environment. Making sense of that data required stitching those passes together later, often by sending large volumes of raw data to the cloud for post-processing.
Dareesoft also found that the agency’s records often did not match reality. Locations were off, attributes were wrong, and some signs were missing entirely.
What Changed
Precise GPS addressed the root of the problem.
By integrating Swift Navigation’s Skylark Precise Positioning Service, the system moved from meter-level accuracy to centimeters. Each detection aligned to the same physical location every time.
With precise GPS, all observations share the same global reference frame. Each detection is anchored to a consistent location as it is captured, across vehicles and across time.
That allows the system to disambiguate detections in real time, at the edge, instead of after the fact. The output is already structured and consistent when it leaves the vehicle.
Within a short period, the system identified 2,496 signs, including 90 damaged and 20 missing.
Turning Fleets Into Mapping Systems
With precise GPS, systems like Dareesoft’s Road Analyzer can be deployed across entire fleets, turning everyday vehicles into continuous mapping sensors.
Buses, service vehicles, and other fleet assets can contribute to a shared, constantly updated view of road infrastructure. Instead of relying on periodic surveys, agencies can keep their maps current through ongoing operations.
In practice, this means map creation and maintenance are no longer separate processes. They become part of daily operations, powered by vehicles that are already on the road.