Question about lena-dual-stripe buildings

[Mattia Rebato]

Hi everybody, 

I'm a ns-3 newbie and this is my first post in this group (hope this post follows all the rules of the group!!). 

I'm running a simulation with the "lena-dual-stripe" example (without any changes in the code of the release 2.23) and I don't understand why in my simulation result (attachment 1) that path loss model is HybridBuildingsPropagationLossModel in all the area. This is different comparing the result on the LENA tutorial (attachment 2) where it's possible to see different path loss model depending if there are buildings. Why the SINR is not attenuated after the signal passes through the buildings we put. In my simulation, it’s like buildings don’t affect the whole scenario.

Moreover the code has been run with the following parameters:

--generateRem=true --epc=false --nMacroEnbSites=1 --nFloors=1 --nBlocks=5 --nApartmentsX=25 --macroEnbTxPowerDbm=55”

It this normal? Or should I obtain a different REM map? 

Thanks in advance!!!

[Tommaso Pecorella]

Hi,

it seems to be normal. The actual REM depends on the simulation params, and if you don't use the same params that the authors used for the documentation, you'll have a different map.

Unfortunately, I have no idea of what the exact params are.

Have fun,

T.

[Mattia Rebato]

Thanks Tommaso for your answer.

Anyway, I don't understand why my actual REM should be normal, knowing that this REM map is made for z=1.5 and buildings are higher than 1.5 meter. The value of SINR obtained from the only macro eNode shouldn’t be attenuated passing through the walls of the buildings? 

Shouldn't be smaller inside and after the buildings? 

Here it seems that buildings are not considered in the computation of the REM (not the same for tutorial map).

Hope you'll understand my question, maybe it's normal like that and I didn't understand how the script works!

I think here the problem is the same.

Regards

Mattia

[Tommaso Pecorella]

Hi,

no idea. Ask Marco Miozzo or Nicola Baldo... they're the best to analyze the issue.

Have fun,

T.

[Biljana Bojović]

Which ns3 version are you using? Did you try with different antenna parameters? When you change the ¨z¨ does it change the REM map?

Biljana

On Monday, June 29, 2015 at 3:00:01 PM UTC+2, Mattia Rebato wrote:

[Nicola Baldo]

Hi Mattia,

you can verify that indeed wall penetration losses are considered by running a simulation with no UEs (so that RemSpectrumPhys will be the only receiving objects) and enabling logging for MultiModelSpectrumChannel and HybridBuildingsPropagationLossModel, like this:

NS_LOG=MultiModelSpectrumChannel:HybridBuildingsPropagationLossModel  ./waf --run lena-dual-stripe --command="%s --generateRem=true --epc=false --nMacroEnbSites=1 --nFloors=1 --nBlocks=5 --nApartmentsX=25 --ns3::Building::ExternalWallsType=ConcreteWithoutWindows --macroUeDensity=0 --homeUesHomeEnbRatio=0 --homeEnbActivationRatio=0"  &> log.txt

Now, the non-trivial fact to understand is that wall penetration loss is indeed being considered, but at the same time your REM is correct. Why? Because in the region that you are plotting and with the parameters that you are considering for the macro eNBs the SINR is dominated by interference from the other macro cell sectors. Note that in your REM the SINR only depends on the angle w.r.t. the macro cell site, but not on the distance from the macro cell site; this is also evident from the fact that the max SINR is 20dB, which corresponds to the max back lobe attenuation of the antenna pattern used by lena-dual-stripe (which in turns corresponds to typical 3GPP sim parameters, e.g., those in R4-092042).  The signals from all macro cell sectors are attenuated by walls in the same way, so when calculating SINR (which is ~= SIR in this case) the wall penetration loss cancels out.

As a counterproof, you can indeed see the effect of wall penetration when you are in a noise-limited scenario, e.g., by considering a lower macro eNB TX power:

./waf --run lena-dual-stripe --command="%s --generateRem=true --epc=false --nMacroEnbSites=1 --nFloors=1 --nBlocks=5 --nApartmentsX=25 --ns3::Building::ExternalWallsType=ConcreteWithoutWindows --macroUeDensity=0 --homeUesHomeEnbRatio=0 --homeEnbActivationRatio=0 --macroEnbTxPowerDbm=10"

which results in the attached REM. Note that we still see the previous behavior for part of the building that is closer to the macro site.

[Mattia Rebato]

Hi Biljana and hi Nicola,

 

first of all, thank you for your answers and explanations. 

I'll consider this in my next simulations. 

Regards,

Mattia