#Wireless/Mobile Networks Project ##Measurement-driven model of radio propagation In this project we built a measurement-driven model of radio propagation. The basic idea is to measure the received power (generally called received signal strength or RSS) at different distances from a wireless transmitter and then fit a model using these measurements.
####Measurements:
The first step is to collect a number of RSS measurements tagged with distance of the transmitter. We used WiFi client devices (laptop) and collected measurements from WiFi AP (for WiFi signals) with a known location. "WireShark packet capture tool" is used to collect RSS for different WiFi Aps on different computer platforms. To find distance, measurement locations are recorded. Locations are determined via a phone’s GPS and using online satellite maps.
####Analysis:
######Following are general methodology for fitting a propagation model to the data.
-
Express RSS in dBm and distance d in meters. Plotted the measurement samples in two scatterplots – RSS vs. d and RSS vs log10(d).
-
We know (log-distance path loss and log-normal shadowing) that d RSS(d) = RSS(d0) − 10α log10 d + Xσ, where α is the path loss exponent, d0 is a reference distance and Xσ is a shadowing component repre- sented by a Gaussian (normal) random variable with zero mean and standard deviation σ. According to the above expression, RSS vs. log10(d) plot should be linear ignoring the random shadowing com- ponent.
-
Used linear regression to find the relationship between RSS and log10(d). The slope of the regression line will give you the value of α. Compute α. (Spreadsheet program- Excel has built-in functions which is used to compute linear regression) Plotted the regression line on the RSS vs. log10(d) scatterplot. Also, plotted the corresponding RSS vs. d line on the RSS vs d scatterplot.
-
Shown the distribution of shadowing component. This is simply the deviation of the measurement samples from the regression line. To do this, computed the algebraic difference between the measured RSS for each d and the corresponding RSS on the regression line. Plot a histogram of these differences. For the histogram we chose a suitable bin size (50 dBs). Also computed the standard deviation σ of these differences. (Using Excel to compute the standard deviation.)
-
In RSS vs d plot, Plotted 4 more lines: RSS ±σ and RSS ±2σ away from the regression line. These additional lines visually show how noisy the shadowing component is.