MF-01·06 · Signals & Comms

Bodywake

coming soon

Detecting a person walking through a house from the shadow their body casts on ambient WiFi. No camera, no wearable, no app on their phone. Just one laptop that never moves.

Full screenComing soon

Part no.	MF-01·06
Room	01 · Signals & Comms
Status	coming soon
Does	one stationary scanner · z-scores, not raw dBm · room-level, honestly
Runs as	A live simulator: calibrate the empty house, then walk through it
Made with	Python · FastAPI · WebSockets · netsh

Key plan · room 01, Signals & Comms

Fig. 1 · Bodywake, on the bench: power it on to run it hereMF-01·06

01Why it exists

A body is 60% water

Water absorbs 2.4 and 5 GHz. Walking through a room shifts RSSI by one to three dBm, and it shifts each access point by a different amount depending on what's between it and the receiver. That difference is the signal. Bodywake reads what the house was already broadcasting.

02Parts list

01 · BaselineScan an empty house for 30 seconds; a stable mean and spread for every BSSID.
02 · PerturbationEvery live reading becomes a z-score against that baseline. The deviation is the motion signal; raw dBm never is.
03 · Zone inferenceEach room has a signature: which BSSIDs deviate, and in which direction. Live deviations are matched against the library; past 2.5σ the zone trips.
04 · TrackZone transitions over time become a path. Room-level resolution, no better, and the code doesn't pretend otherwise.

03Construction

The rewrite

The first architecture solved the wrong problem: it fingerprinted the scanner's own position (which room is the laptop in?), a solved and useless question when the laptop never moves. The corrected model keeps the scanner fixed and detects someone else by how their body perturbs the ambient field.

Bugs that cost real time

An asyncio.Lock() built outside the event loop, a deprecated startup hook, and (the dumbest) a Windows launcher that died on a folder path with spaces and a trailing backslash. Windows scans go through netsh, which caps the poll rate and quietly caches; that ceiling, not the physics, limits time resolution.

PythonFastAPIWebSocketsnetsh

04Tolerances

Presence and motion are reliable; room-level zoning works after calibration. Sub-room position, counting people, and seeing through more than a couple of walls without a second access point don't. The simulator's physics are simplified; the method is the real one. The upgrade paths are CSI (sensitive enough for breathing) and 802.11mc round-trip timing.

what it is not, said plainly

05Inscription

Ambient RF is already everywhere in the house, already carrying the information.