Wifi-Positioning-System

This project aims at getting the coordinates of a receiver in an area having 3 access points or Wifi Routers which act as transmitters. This consists of 3 steps: Firstly getting the received signal strength in dbm by the receiver from all the 3 transmitters, Secondly getting the distance from the 3 access points through the received signal strength from access points using Logarithmic Distance Path Loss Model which describes the path loss by a signal as a function of distance, and lastly calculating the coordinates of receiver if coordinates of transmitter and distance from them is known.

Prerequisites :

The project is made using Python, so requires a knowledge of Python. Other libraries used are as follows: -

-> subprocess
-> numpy
-> time
-> openpyxl
-> math

Theory:

The project uses subprocess library to get output of Command Prompt command 'netsh wlan show network mode=Bssid' which gives a list of the wifi signals present near the receiver and their related information from which we can get the received signal strength in %. To convert it into dbm I used a scale of -50 dbm to -100 dbm. -50dbm represents 100% signal strength while -100 dbm being the least received signal strength represents 0% and used linear interpolation to get received signal strength in dbm from %. In the next to calculate the distance from 3 access points from their corresponding signal strength I used Logarithmic Path Loss Model which is as follows: -

Pl(d) = Pl(d0) + 10*n*log(d/d0) + X
where Pl(d) -> path loss at a distance d from transmitter
n -> attenuation factor
d0 -> reference distance from transmitter usually taken 1m
Pl(d0) -> path loss at distance d0 from transmitter
X -> Guassian Random Variable with zero mean describing attenuation caused by flat fading.

Pl(d) = PT - PR
where PT -> transmitted power in dbm
PR -> received power in dbm

Therfore Pr can be effectively written as
PR = Pl(d0) - 10*n*log(d/d0)
now here Pl(d0) is constant which takes care of transmitting power of transmitters, 
path loss at distance d0 and Guassian Random Variable

So distance d can be given as : -
d = d0*10^((Pl(d0)-Pl(d))/10*n)

Now once the distances are calculated and if we know the coordinates of the access points or the transmitters then we can calculate the coordinates of receiver by distance formula. In this case of 3 access points we will get 3 equations corrsponding to distances from 3 access points. These equations will be containing square terms so subtract the last equation from the rest two to get rid of sqaure terms and write the resulting 2 equations in the form matrix multiplication of the form Ax = b and use least square algorithm to get x which will give the coordinates of receiver.

Contents: -

• parameters.py : python code for getting the parameters Pl(d0) and n. Through this file you can take 20 readings of recieved signal strength at a point and then take average of them to get average recieved signal strength at that point and this you can done at 10 points to get the approximate values of Pl(d0) and n for that environment. If transmitting power of wifi routers is different then you need to run this file for every wifi router as Pl(d0) now also describes transmitting power of transmitters.
• access_points.py : this is main code that is responsible for getting the coordinates of the receiver. Remember that you need to manually feed the values of Pl(d0) and n for different routers as well as the coordinates of access points.
• neural_network.py, neural_network_approach.py : Apart from the above solution neural network can also be used for getting the coordinates if received signal strength is the input to the neural network. But this requires a lot of data for training the neural network and getting a good accuracy so is currently in development phase.