Video Summary - Can your WiFi router see through walls?

Scientific concepts / discoveries / phenomena

Wi‑Fi-based sensing of human presence and motion
- The video describes using Wi‑Fi signals to detect a person’s position and movement in a room.
Channel State Information (CSI) and subcarriers
- Wi‑Fi transmits data over dozens of parallel subcarriers.
- For each subcarrier, the receiver can measure CSI, including:
  - phase
  - amplitude
  - arrival timing/strength (signal characteristics captured by CSI)
Human body as a signal perturbation source
- The body is stated to be ~60% water.
- Water absorbs and reflects Wi‑Fi signals, particularly at 2.4 GHz and 5 GHz.
- When a person moves, they distort the subcarriers, producing CSI changes.
Machine learning for converting CSI “noise” into human-readable information
- Raw CSI disturbances are treated as a form of signal “noise” that is meaningful.
- A neural network is trained on millions of CSI patterns paired with known body positions.
- The trained model maps CSI patterns → inferred human position.
Through-wall capability and material penetration
- The video claims radio waves can pass through materials (e.g., drywall, wood, concrete) up to ~30 cm, enabling sensing through walls.
Threat model (privacy/security implication)
- It suggests that if an attacker has access to a person’s Wi‑Fi router (or can intercept/measure the relevant signals), they could potentially use this method to spy indoors.

Use two ESP32 boards
Flash open-source firmware onto the ESP32 devices
Run processing/inference software on a computer
Place one ESP32 near the person (and the second at another location/position)
Record Wi‑Fi signals and CSI
Use a trained neural network to infer body presence/position from CSI patterns

No specific researchers or organizations are named in the provided subtitles.