Design Fleet Map Data Collection
Coordinate a vehicle fleet that captures, uploads, validates, and publishes fresh map observations.
Problem statement
The source asks for a vehicle fleet that collects mapping information. For this exercise, assume fleet vehicles receive geographic collection missions, capture timestamped sensor segments, buffer them through intermittent connectivity, and upload immutable encrypted bundles.
The backend tracks coverage, validates integrity and calibration, deduplicates overlapping observations, and promotes reviewed map updates into versioned regional releases. Collection vehicles must not consume an unreviewed upload as authoritative map data.
Route planning and vehicle-control algorithms, human labeling tools, and the internal map-fusion model are outside scope.
What to build
- Operators can create coverage missions and assign compatible geographic segments to available vehicles.
- Vehicles can resumably upload immutable sensor bundles with calibration, location, time, and integrity metadata.
- Operators can inspect collection freshness, gaps, duplicates, validation state, and vehicle progress.
- Validated observations can feed an auditable, versioned regional map release without changing prior releases.
At scale
- For this exercise, assume 25,000 vehicles, each producing up to 500 GB of selected data per day with long offline periods.
- For this exercise, track coverage at road-segment and time-window granularity for 100 million segments.
- For this exercise, resume interrupted uploads at chunk boundaries and make newly accepted metadata visible to operations within five minutes.
- For this exercise, retain raw selected observations for 30 days and immutable promoted evidence for the lifetime of a map release.