How It Works

From WiFi packets to AI detection

A six-stage pipeline turns invisible radio reflections into actionable human detection.

Step 01
Capture

ESP32 or CSI-supported WiFi hardware collects Channel State Information packets.

Step 02
Reflect

Human movement changes how WiFi signals reflect, scatter, and arrive at the receiver.

Step 03
Process

RuView processes CSI amplitude and phase variations to extract motion features.

Step 04
Infer

AI models analyze patterns for presence, gestures, motion, breathing, and heartbeat.

Step 05
Detect

The system classifies events: occupancy, movement, vitals, and posture.

Step 06
Visualize

Results stream into real-time dashboards or downstream automations.

What is WiFi CSI?

Channel State Information is per-subcarrier amplitude and phase data extracted from WiFi packets. It exposes how the radio channel changes over time — including from human movement.

How WiFi Signals Detect Movement

When a person moves in the channel, multipath reflections shift. AI models learn the signature of those shifts to classify what's happening.

Signal Reflection Explained

Bodies are mostly water and reflect 2.4/5 GHz energy. The receiver sees subtle changes in delay, phase, and amplitude across subcarriers.

CSI Data Collection

Firmware on ESP32-class chips exposes CSI for every received frame, giving rich raw input for machine learning.

AI Inference Pipeline

Preprocessing, feature extraction, and a lightweight neural network produce predictions — usually under 100ms end to end.

Edge Processing

Everything runs locally on the device or a nearby host. No cloud round-trip is required.

Signal Visualization

Spectrograms, occupancy timelines, and per-subcarrier heatmaps make CSI interpretable to humans.

Limitations and Realistic Expectations

WiFi sensing is probabilistic. Through-wall metal, dense furniture, and interference all affect accuracy. Treat outputs as estimates.