From WiFi packets to human signals
RuView's pipeline turns commodity WiFi traffic into structured human detection events using a Rust signal engine and edge neural inference.
WiFi CSI
Channel State Information describes how each WiFi sub-carrier was attenuated and phase-shifted by the environment. RuView captures CSI for every packet.
Signal reflection
Walls, furniture and bodies reflect, absorb and bend 2.4 / 5 GHz signals. These distortions are the raw material of WiFi sensing.
ESP32 data collection
ESP32-S3 / C5 nodes stream raw CSI over UDP or TCP to the RuView engine — typically thousands of packets per second.
Neural processing
The Rust engine denoises, aligns and reduces dimensionality (PCA), then feeds time-series tensors into compact neural models.
AI inference
Models classify presence, motion type, pose, fall events and extract vitals like breathing and heart rate.
Signal visualization
Live dashboards stream CSI heatmaps, motion timelines and event logs over WebSocket / MQTT.
Every visible pulse is a WiFi packet being analyzed in real time by the RuView edge AI engine.
The full RuView signal pipeline
ESP32 (CSI fw)
│ raw subcarriers · phase · amplitude
▼
Signal cleanup ─────▶ Calibration ─────▶ PCA / feature extraction
│
▼
Neural inference
│
▼
┌───────────────┴───────────────┐
▼ ▼ ▼
Presence Motion Vitals
│ │ │
└──────── Event bus (WS / MQTT) ┘