Real dashcam · one visit,
Imagery: NJ Office of GIS, 2020
Inbound (gate → yard)
In the yard (container handling)
Outbound (yard → exit)
Simulated truck stopped at a booth or in a queue
The real truck in the dashcam
What's in the map
Whole visit (minutes)
| Stretch | Real p10 / med / p90 | Sim p10 / med / p90 | KS D |
|---|
Yard time runs from leaving canopy E to reaching the out-portal. Yard dwell is fitted to our trucks' real visits; the drive through the yard comes from the map.
Gate lines
| Line | Lanes | Real time from previous line (median s) |
|---|
Lane counts come from the booth spacing on New Jersey's 2020 state orthophotography (±2). Portal counts were taken by eye on the same imagery.
Real vs simulated gate times (seconds)
| Segment | Real p10 / med / p90 | Sim p10 / med / p90 | KS D |
|---|
Simulated at the fitting volume. KS D: 0 means identical distributions; below 0.25 is a close match for these sample sizes.
When does the gate jam?
Limits
- Real data: our drayage trucks, Jul–Oct 2026, daytime only. Every path and timing comes from them.
- Truck volume isn't observable from our fleet. The volume sweep shows what the real wait times would allow; pinning it needs gate transaction counts.
- The yard is one waiting area along the main aisle where our trucks waited most. Real trucks are served at stacks across the whole yard.
- Booth service is fitted as a stop at each canopy line. Real trucks creep forward in queues; the sim stops them in one go.
- Lanes are per canopy, not per painted stripe. The open plaza between canopies is modelled as wide fan roads.
- The layout is traced on New Jersey's 2020 state orthophotography (NJ Office of GIS). Changes to the booths since that flight aren't shown.
- Faces in the dashcam clips are blurred. The camera stops uploading while a truck is parked, so the container loading itself isn't on video.