how to obtain pathgains of winner2 channel

the channel is generated by the following code

BSAA  = winner2.AntennaArray('UCA', 1, 0.02);  % UCA-8 array for BS
MSAA1 = winner2.AntennaArray('ULA', 1, 0.01);  % ULA-2 array for MS

MSIdx = [2]; 
BSIdx = {1}; 
numLinks = 2; 
rndSeed = 5;
cfgLayout = winner2.layoutparset(MSIdx,BSIdx, numLinks,[BSAA,MSAA1],[],rndSeed);

cfgLayout.Pairing = [1 ; 2 ];  % Index in cfgLayout.Stations
cfgLayout.ScenarioVector = [6]; 

numBSSect = sum(cfgLayout.NofSect);
numMS = length(MSIdx);


cfgLayout.Stations(2).Velocity=[1.6;0;0];
cfgLayout.Stations(1).Pos(1:2) = [50, 150];
cfgLayout.Stations(2).Pos(1:2) = [10, 180];  % 8 meters away from BS

BSPos = cell2mat({cfgLayout.Stations(1:numBSSect).Pos});
MSPos = cell2mat({cfgLayout.Stations(numBSSect+1:end).Pos});

frameLen = 64; % Number of samples to be generated

cfgWim = winner2.wimparset;
cfgWim.NumTimeSamples      = 80;
cfgWim.IntraClusterDsUsed  = 'yes'; 
cfgWim.CenterFrequency     = 2.76e9;
cfgWim.UniformTimeSampling = 'yes'; 
cfgWim.ShadowingModelUsed  = 'yes';  
cfgWim.PathLossModelUsed   = 'no';  
cfgWim.UseManualPropCondition = 'yes';
cfgWim.RandomSeed          = 31415927; 
cfgWim.NormalizeChannelOutputs='false';

downWINNERChan = comm.WINNER2Channel(cfgWim,cfgLayout);

and the input data is

sourceData = randi([0 1],(80,1);

then the pathgains and output are obainted

[ChanOut, pathGains] = downWINNERChan(sourceData);
pathGains1=squeeze(cell2mat(pathGains));
pathGains1=sum(pathGains)';
rx_data=cell2mat(ChanOut);
tx_data_estimate=rx_data./pathGains;

I wonder why tx_data_estimate is different from sourceData?

Answer 1

[Yukthi S]

Hello

I see that you are observing some discrepancies between ‘tx_data_estimate’ and ‘sourceData’.

Firstly, it is important to understand that WINNER II channel model simulates a complex wireless communication environment including effects like multipath fading, shadowing, noise and antenna patterns. These effects inherently modify the signal in a way that the received signal (rx_data) cannot be directly divided by the path gains (pathGains) to perfectly reconstruct the original transmitted data (sourceData). The path gains represent the complex channel effects on the signal and simply dividing the received signal by these gains doesn't fully explain all the changes the signal goes through as it travels through the channel.

But, if you still want to perform this operation to calculate estimate of transmitted signal, there are some errors in the code snippet you provided which need to be addressed.

sourceData = randi([0 1],80,1);
%instead of sourceData = randi([0 1],(80,1);

The line "pathGains1=squeeze(cell2mat(pathGains));" attempts to convert pathGains from a cell array to a matrix, which is correct in concept. However, the subsequent operation "pathGains1=sum(pathGains)';" seems to be an attempt to sum the path gains but is incorrectly applied to the unprocessed "pathGains" variable rather than "pathGains1". The corrected approach would be:

[ChanOut, pathGains] = downWINNERChan(sourceData);
pathGains1=squeeze(cell2mat(pathGains));
pathGains2=sum(pathGains1)';
rx_data=cell2mat(ChanOut);
tx_data_estimate=rx_data./pathGains2;

I hope this addresses your query!