Start a Project

Case Study / Smart Infrastructure

A parking network
that manages itself.

For a multi-location parking operator based in the USA, EdgeWeb built a real-time system that detects which bays are free, prices every hour dynamically against live demand, and opens the barrier without a human at the booth — across every lot, from one platform.

The Problem & The Approach

The problem

Every lot ran on its own logic: attendants collecting cash at a boom barrier, one flat rate no matter how full the lot was, and no shared view across locations. Prime spots sat empty while overflow lots turned cars away, entry queues backed up onto the street at peak hours, and underpaid or unpaid exits quietly leaked revenue every single day.

The approach

Replace the attendant's judgment call with a system that actually knows, in real time, how full every lot is — then let that number drive the price, the barrier, and the bill automatically. One platform now runs pricing, entry, exit and billing across every location the operator has.

Core System Logic

Six pieces, one continuous loop.

Real-Time Occupancy Engine

An ultrasonic sensor over every bay reports state changes over MQTT the instant a car arrives or leaves. A central occupancy service aggregates that into a per-lot, per-zone live count — the single source of truth everything else reads from.

Reservation & Locking

Reserving a bay through the app places a short-lived lock on that exact bay — typically 15 minutes. If the driver doesn't show, the lock expires and the bay silently returns to the available pool instead of sitting reserved-but-empty.

Dynamic Pricing Algorithm

The hourly rate is the base rate multiplied by a demand tier computed from current occupancy — recalculated the moment a bay's status changes, not on a fixed schedule. The live simulator below runs the exact same tiers.

Automated Entry & Exit (ANPR)

A camera at each barrier reads the plate on approach. If it matches an active reservation or a stored payment profile, the barrier lifts with no ticket, no button, no attendant — on the way in or out.

Usage-Based Billing

Exit closes the session, calculates duration × the rate that applied at each point in the stay, and charges the saved payment method automatically — or presents a pay-on-exit screen if none is on file. A receipt goes out either way.

Multi-Location Ops Dashboard

Every lot's occupancy, revenue and sensor health on one live map, with an alert the moment a sensor goes offline or a bay reads "occupied" for implausibly long — usually a dead sensor, occasionally an abandoned vehicle worth checking on.

Try It

See the pricing engine live.

This runs the same tiered logic as the production system. Click an open bay to reserve it, or turn on live traffic to watch occupancy — and the rate — move on their own.

Live pricing engine

$3.00/ hour · Base rate
Occupancy36%
Available — click to reserveOccupied (live sensor)Reserved by you — click to release

Architecture

Chosen for real-time load, not a demo.

IoT ultrasonic sensorsMQTT brokerANPR camera integrationNode.js (event-driven)PostgreSQLRedis (reservation locks)React admin dashboardReact Native mobile appStripeAWS / cloud hosting

Outcomes

42%
Faster average time to find and enter a bay
19%
Revenue increase from demand-based pricing
0
Attendants required for entry or exit
99.9%
Platform uptime across all locations

FAQ

Does this replace our existing barriers and cameras?

Not necessarily — most ANPR cameras and boom barriers speak standard protocols the platform can integrate with directly. We only replace hardware where it can't be integrated or is past end of life.

Can pricing rules differ per lot?

Yes. Base rate, demand tiers, caps and even time-of-day rules are configured per location from the admin dashboard, not hardcoded — a downtown lot and an airport overflow lot rarely want the same curve.

What happens if a sensor or camera goes offline?

The dashboard flags it immediately, and that bay falls back to a manual-override state rather than silently reporting stale data. Pricing for the rest of the lot keeps working off the sensors that are still online.

How long did a project like this take?

Rolling out the platform to a first pilot lot typically takes 8–12 weeks; each additional location after that is mostly a hardware install and configuration exercise, not new software.

Have a project in mind?Start a Conversation →