##Dependence##

• numpy
• scipy
• sklearn
• pygame

##Introduction## This is a wlan positioning solution framework. It is proposed on the basis of machine learning and therefore composed of two phases:

• offline phase
  1. process raw data
  2. train model
• online phase
  1. launch positioning server
  2. predict user's position This frameword is implemented as client-server mode.

##Usage##

1. data first of all, make a new directory and put your offline-phase data into 'raw_data/[your_dataset]' put recieved signal strength(RSS) files into 'raw_data/[your_dataset]/rss' put map infomations into 'raw_data/[your_dataset]/map' and put test data into 'raw_data/[your_dataset]/test'

2. run framework

   ./go.py -d [dataset] -a [alg]
   ./go.py -d [dataset] -a [alg] env
   ./go.py -d [dataset] -a [alg] offline
   ./go.py -d [dataset] -a [alg] online

   • options:
     • -d: specify your dataset in raw_data
     • -a: specify your machine learning algorithm in alg
   • arguments:
     • env: build an environment for further operation
     • offline: execute offline operation, should execute env first
     • online: execute online operation, should execute env and offline first
     • test: should be used seperately
     • nothing specified: will execute env, offline and online sequentially

3. run PC client

   ./pc_client.py

4. run phone client use app in phone_client

5. useful tools

   • rss collector in phone_collector directory
   • map information collector in offline_tool directory

##BUG REPORT##

• zchgeek@gmail.com
• wchgeek@gmail.com