Realistic Solutions for Real People

Data Matrix

iModel

iModel, the latest software created by Data Matrix, produces accurately received signal strengths from various antennas.  It calculates received signal strengths by using previously surveyed data.  iModel predicts a path loss exponent inside and outside an azimuth.  This technique provides iModel with an upper hand against competitive software.

Data Matrix was given a chance to predict signal strength for 4 antennas located on the campus of Georgia Institute of Technology.  They were also given data from four other antennas.  Engineers at Data Matrix collected data from four antennas named Cell Info A, Cell Info B, Cell Info C, and Cell Info D. Coverage maps of A-D can be seen as the first item in the Coverage Maps page.  The gathered data consisted of power received from a 203 x 203 matrix with each point representing 10 meters.  The data from the four given antennas was used to find power received for four mysterious antennas.  The power received was used to calculate path loss for each r using the following equation:

 

[1]

Assumptions: R_ref = 1 m and G_T = 0 dBi

Source Code 1 was used to calculate r (separation distance) for all the points that had a power received of less than 0.  The P_R and the coordinates values are also found from this code.  This finding of r was useful when finding path loss at each r.

Blank matrices were created for the finding of path loss at r.  Engineers designed an algorithm to find the path loss inside and outside an azimuth.  The azimuth for the antennas represents the direction the antenna is pointing.

The path loss equation is used in conjunction with the following equation to find the path loss exponent:

 

 

 

[2]

The engineers were able to find data inside the azimuth using a loop. Source Code 2 can be viewed to see the loop used to get inside the azimuth.  This was done because if you know the azimuth of the antenna, you can solve for the limits inside the azimuth.  The engineers changed the loop depending on the azimuth.  This loop was used to find two path loss exponents, because if you were not in the loop then you were obviously not in the azimuth.  This was done by adding an else statement.  Engineers found two values for each variable needed.  One set of variables would be inside and other would be outside the azimuth.  Source Code 3 shows the two variables.  Using the two path loss exponents at certain locations, engineers predicted more path loss exponents to produce iModel.  iModel uses the predicted path loss exponents in an algorithm to calculate power received from a mixture of antennas.  The predicted path loss can be used to find power received by using the following equation:

 

[3]

After finding the path loss exponents for each antenna, engineers put n back into the budget link equation to find the signal strength for antennas A-D.  Power received data from antennas A-D using this model can be seen in the second item under Coverage Maps.  The table below shows the mean and standard deviation of the model for each antennas. 

 

 

 

 

 

The mean was found to be 3.4 dB and sigma was 11.0 dB.  With this the Raw Performance Score (RPS) was calculated to be 11.2. The fourth item in Source Code Examples page contains the screen shot of running the CompareMaps function.  Overall, Data Matrix was pleased with their results.  

The path loss exponents used to predict power received were estimates from experiments with antennas A-D.  Since antenna A and D were in the same location as antenna E, engineers averaged the path loss exponents for inside and outside the azimuth from both antennas to get a new inside and outside path loss exponents for E.  This was the same method used for predicting path loss exponents for antennas F-H, because antenna B was in the same location as antenna F and G and antenna C was in the same location as H.  Using this estimation may cause the data to be slightly inaccurate.  With the new path loss exponents, engineers used the Budget Link Equation (equation 3) to calculate the signal strength for the mysterious antennas.    

Source Code 4 shows the Matlab code using an estimated path loss exponent to predict the power received for the mysterious antennas.  The azimuth was set for each particular antenna for E-H.  The engineers also produced coverage maps for E-H after running.  These maps show power received for the mysterious antennas and can be viewed at Prediction Maps for E-H.

We at Data Matrix, believe this strategy can accurately predict and map signal strength for various locations.  We will prosper when our results are compared to the real values of the mysterious antennas because of our technique, which considers the inside and outside of the azimuth.  The method is more accurate than others.  The technique uses surveyed data which produces an inside and outside azimuth for each antenna.  With that, we estimate two path loss exponents which is used to predict signal strength.     

Data Matrix will help lead the world to produce better wireless networks by being able to predict signal strength over a given area.  Lets take for example the predicament with cellular companies.  Cell phone users hate when their calls are dropped or when they lose signal.  With these coverage maps, companies will be able to accurately tell where more antennas are needed, so people don’t drop calls.  This will help increase the satisfaction rating from the consumer.  The increase in satisfaction will lead to more customers and revenue for companies.  The information that Data Matrix provides will also help make the wireless network in a company run more efficiently.  The coverage maps can show if two antennas overlap significantly.  This would be inefficient to have a significant overlapping, so the coverage maps again will greatly benefit a company.