##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
- process raw data
- train model
- online phase
- launch positioning server
- predict user's position This frameword is implemented as client-server mode.
##Usage##
-
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'
-
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
- options:
-
run PC client
./pc_client.py
-
run phone client use app in phone_client
-
useful tools
- rss collector in phone_collector directory
- map information collector in offline_tool directory
##BUG REPORT##