Live demo
AdvancedTry it: count event foot traffic with WiFi sensing
This is the working version of the tutorial: a simulated trade show floor with a WiFi doorway counter, two booth zones, and a live zone flow graph. Your browser cannot read real WiFi radio data, so every signal here is simulated. Doorway estimates use threshold crossings and trigger pairing. Booth dwell, occupancy, and zone flows use the animation's known attendee positions and IDs, not recovered radio measurements. The physical kit uses ESP32 boards that cost about $12 each. Simulation runs at 3x real time.
Simulating…
Bluetooth: two halves, one honest
The beacons you deploy (B-01 at the entrance, B-02 and B-03 in the booths) broadcast stable IDs: reliable zone presence. Attendee phones randomize the address they show while probing, so passive sniffing cannot track a person across the day. Watch the phone addresses rotate below.
| Device | Address seen | Stable? |
|---|
Doorway counter
Links A and B sit about one meter apart. A disturbance spike crossing the threshold is a trigger; A-then-B means entering, B-then-A means exiting. Toggle rush mode and watch bunched groups cross as one trigger.
Entries: 0 Exits: 0 Simulated inside now (ground truth): 0
Booth zones and dwell time (simulation ground truth)
| Booth | Present | Visits | Median dwell | Longest dwell |
|---|
Recent visits (entry to exit):
Zone flow graph (simulation ground truth)
Zones are nodes, movements are edges from the animation's exact positions. This illustrates a possible report, not measured anonymous visit tracking.
Doorway estimates from simulated signals
| Door | Entries | Exits | Net |
|---|---|---|---|
| Main entrance | 0 | 0 | 0 |
Starter kit
The starter kit holds the aggregator sketch (Python: UDP listener, trigger pairing, CSV log), firmware pointers for the ESP32 boards, the placement guide, the calibration checklist, and the disclosure template.
Want the files behind this lab, sample dataset included? Email ai@trihelix.co and we will send them your way.