Skip to content
Planta Smart Homes logo
Buy
Business Solutions

Filmed on site · Rock in Rio Lisboa · Parque Tejo

Three films. One festival, reading itself.

Every clip below was recorded at the deployment. Nothing is a render, and no film is repeated.

Film 01 · The ground truth

80,000 people a day

The environment Planta had to read: 145,000 m² of festival, eight WC clusters, 1,137 cabins, and a crowd that moves in waves after every headline set.

Film 02 · The routing

Sending the crowd to the empty side

Sensors at each portal count entries and exits. When one cluster fills, the screens and the app name the emptiest alternative — by number and by percentage.

Film 03 · Smart City of Rock

The full system, end to end

Sensor to screen in under three seconds, entirely anonymous: no cameras storing faces, no identity, no tracking — only how many people crossed a line.

Selected by Rock in Rio among thousands of international startups

Rock in Rio screened the world. Planta shipped in the field.

  • One team, the whole stack

    Planta built the sensors, enclosures, radio network, model, app and venue screens — no integrator in between.

  • Live, not a pilot

    Four nights, 8 WC clusters, 1,137 verified cabins, across 145,000 m² of open festival ground.

  • Anonymous by construction

    The sensors count crossings, not people. No faces, no identity, no device tracking — aggregate numbers only.

  • It recommends, not reports

    Instead of a dashboard, the system names the next best WC: number, occupancy, walking direction.

  • Calibrated to real crowd waves

    Behaviour differs between light hours, growth, peak and the post-concert surge; the model is tuned to each.

  • Seen by ~50,000 people

    In the app and on the on-site screens during the festival — measured, not estimated.

The problem

At large festivals, queues become invisible friction.

People lose time. Comfort decreases. Movement slows. Revenue opportunities move more slowly. WC pressure is one of the most important flow problems in festival operations.

  • queue pressure
  • waiting time
  • movement friction
  • operational blind spots
  • value loss from waiting

Validated internal estimate

~€1M / day

Potential recovered value for large festival flow improvement.

Methodology:
flow & value model
Source:
internal calculation
Validation:
partner / team
Date:
2026

How it worked

Signals become patterns. Patterns become flow intelligence.

  1. 1

    Sense

    At every WC portal, a depth camera and an infrared beam watch the doorway line — not the people, only the crossing.

  2. 2

    Read

    Each crossing is signed: in or out. Entries minus exits, per cluster, per gender side, second by second.

  3. 3

    Verify

    Vision and infrared must agree before a count is accepted, so groups, hesitation and doorway loitering do not inflate occupancy.

  4. 4

    Guide

    Occupancy becomes a decision: the screens and app name the emptiest nearby cluster with its percentage and walking direction.

  5. 5

    Improve

    Operations see pressure building minutes before a queue forms, and can open, clean or staff a cluster ahead of the wave.

Technology

Physical AI infrastructure for live festival flow.

Planta Smart Homes combines sensors, connectivity, edge logic, and software interfaces into one operational flow intelligence system.

  • Vision sensor

    Luxonis OAK-D / OAK-4

    Depth-aware counting at the WC portal; depth only, no faces stored.

  • E18-D80NK

    Infrared sensors

    Beam-break pairs give direction of travel, fused with the vision count.

  • Edge node

    LilyGo devices

    The counting logic runs on the node itself — no video ever leaves the portal.

  • 868 MHz

    LoRaWAN

    Kilometre-range radio that reaches all 8 clusters without festival Wi-Fi.

  • Algorithm

    Sensor fusion

    Two disagreeing signals become one verified occupancy figure.

  • Field hardware

    Ruggedised rigs

    Enclosures, mounts, wiring and power built for rain, dust and four nights outdoors.

  • Guidance UI

    Venue screens

    Large PT · EN screens: green for free, orange for pressure, alternative on the right.

  • Software layer

    App + WebApp

    Visitor app for finding a free WC; operator webapp for calibration and monitoring.

PerceptionCommunicationFusionInterfaceGuidanceOperations

Live venue screens — real captures

What the crowd actually saw at Rock in Rio Lisboa.

Real screen captures from the deployment — per-WC occupancy, the emptiest alternative called out on the right, and graceful handling of partial sensor data.

8 WCs liveGreen = freeOrange = pressureEmptiest alt on rightBilingual PT · EN

The app experience

Planta Smart Homes connected live physical signals with a usable app and web platform.

Planta Smart Homes mobile app — WC map

Mobile App

The mobile experience helps users find the nearest or most convenient WC based on live availability and flow conditions.

  • nearest WC
  • category filters
  • location-based guidance
  • anonymous flow logic
  • simple mobile UX

Impact

A festival read without watching anyone.

  • Visitors sent to a free WC instead of joining the nearest queue
  • Queue pressure spread across all 8 clusters
  • Operations see build-up before the queue exists
  • 1,137 cabins actually used, not just installed
  • Smoother movement between stages and services
  • Seen by around 50,000 people over four nights
  • Live Physical AI proof at festival scale
  • Sensor → edge → radio → screen → app in one system

What it proved

Sensors, radio, screens and app — one intelligent system.

  • Tailor-made sensing for a real venue

  • Hardware Planta builds and mounts itself

  • App, webapp and screens as one guidance layer

  • Anonymous crowd intelligence at scale

Rock in Rio Lisboa · Parque Tejo

The queue was never the problem. Not knowing was.

Want the same live flow intelligence for your venue, event or building? Planta designs, builds and operates the whole system.

Delivered with

Rock in Rio and Unicorn Factory Lisboa selected Planta through the Liquid Innovation programme; SAPO covered the deployment on site.