The operating system for smart parking.

At the edge of the platform sit the parking spaces themselves. Each space is fitted with an occupancy sensor that detects whether a vehicle is present. Sensor types vary by environment — ultrasonic and radar for garages, magnetic loops embedded in open lots, camera-based detection in multi-level structures — and the platform is deliberately sensor-agnostic.

What matters is the message contract, not the hardware. Every sensor reports a facility identifier, a zone and slot, a timestamp, and a confidence level. Whether the physical layer is simulated, retrofitted, or built in from day one, the platform accepts the same events, which is exactly what makes the architecture future-proof.

Sensors push updates over a connectivity layer rather than being polled. MQTT is the workhorse for facilities with reliable IP networks — lightweight publish/subscribe messaging that moves thousands of occupancy events per minute. LoRaWAN covers the long-range, low-power case, letting battery-powered sensors report from every level of a garage for years without external power.

Webhooks let third-party platforms push events in, and edge gateways buffer and batch messages when connectivity drops, replaying them once the link returns. The result: the digital twin never silently misses an event, even across flaky networks.

Events land in a real-time backend that validates, deduplicates, and stores every state change, attaching an exact timestamp at ingress. From that stream, the digital twin is maintained: a live, queryable model of the facility in which every slot has a current state — occupied, free, or reserved — alongside its geometry and metadata.

Because the twin is a model rather than a static diagram, it supports snapshots in time, answers questions like “what did Level 1 look like at 5 p.m. on a rainy Tuesday?”, and acts as the single source of truth for every view in the system. Multi-tenant isolation keeps each operator’s facilities, channels, and settings strictly separate.

Operators experience the twin through two lenses. A real-time dashboard renders occupancy as it happens — utilization totals, per-slot state mapped onto the facility plan, and live alerts when a level fills up. An analytics layer answers the longer view: where occupancy peaks, where queues form, and whether pricing or layout changes actually moved the needle.

Because every event is time-stamped and stored, historical reporting is exact rather than sampled. Underneath it all, the platform runs as a multi-tenant SaaS: operators manage facilities, users, roles, and alerts from a control plane, and the same instance scales from a single lot to a city-wide portfolio.