Skip to content

Latest commit

 

History

History
47 lines (24 loc) · 867 Bytes

README.md

File metadata and controls

47 lines (24 loc) · 867 Bytes

Localization using WiFi scans and Neural Net

Requirements

  • PyTorch
  • Python-Iwlist

Instructions

Training Phase

A scanning phase has to be performed for each different zone.

That is done by running (should be done with root privileges for better results):

python scanZone.py ZONE_NAME NUMBER_OF_SAMPLE

This shoud generate a different json file for each zone.

Merge the training data into a single dataset

That is done by running:

python mergeDataset.py <zone1.json zone2.json ... >

This creates the raw_dataset.json files

Create support files

Run the following commands:

python getAllAP.py

python getAllZones.py

These generates aplist.json and zonelist.json

Train the neural network

That is done by running:

python trainNet.py

Localize yourself

bash localize.sh