LTE: Channel model calibration in ns-3

[Arturo González]

Hi ns3 community,

I am trying to calibrate an ns-3 LTE deployment for a 3GPP Urban Macro Scenario; Case II (Vehicular A).

• For the propagation pathloss I have used the module LogDistancePathlossModel with the calibrated parameters according to 3GPP's TR 25.996; No mayor problem; so far so good.
  
• For Multipath fading, I will generate new fading traces with the provided .m script and velocity of interest to try on my use cases. I am optimistic there is not "much deal" on the generation on those.
  
• I need to add lognormal shadowing with std dev =8 and 0 mean to the current channel; I know this is somehow possible but I cannot figure out how (help on this appreciated);
  
• Some penetration loss might be as well required to be considered at some point.
  

Regarding the two last points; I understand there is a module called BuildingsPropagationModel which sort of groups the last two in one class. I see it only works with the buildings mobility model; a drawback to me since in the future I would like to supply my own mobility models...
Now the questions,
Is there a way to:

1. Generate log normal shadowing without the BuildingsPropagationModel/BuildingsMobilityModel pair? Say I want to do a link level calibration and want to test just shadowing and shadowing+fading+pathloss
   
2. How does this signal degradation models, shadowing+fading+pathloss+penetration loss are "chained together" such that at an overall effect can be seen at REM output ? --(I am relatively new at ns-3 so I might be skipping something fundamental; my apologies if this is the case)
3. Is there some sort of "channel model helper" that does the latter?

Help and guideance at this stage of my work would be highly appreciated.

Best regards,

Arturo G.

[Nicola Baldo]

Hi Arturo,

On Wednesday, November 21, 2012 6:33:17 PM UTC+1, Arturo González wrote:

1. Generate log normal shadowing without the BuildingsPropagationModel/BuildingsMobilityModel pair? Say I want to do a link level calibration and want to test just shadowing and shadowing+fading+pathloss
   

you could create a new ShadowingPropagationLossModel based on the PropagationLossModel interface, possibly leveraging on part of the code from BuildingPropagationLossModel. If you do, please consider contributing the resulting model to the community, I think it might benefit several people.
 

1. How does this signal degradation models, shadowing+fading+pathloss+penetration loss are "chained together" such that at an overall effect can be seen at REM output ? --(I am relatively new at ns-3 so I might be skipping something fundamental; my apologies if this is the case)

PropagationLossModel already supports chaining, and REMs evaluate the overall loss, including chaining if you do it. You just need to setup the chaining properly. Possibly some minor modification to LteHelper might be needed.

1. Is there some sort of "channel model helper" that does the latter?

For doing exactly what you described, no.

Regards,

Nicola


 

[Arturo González]

Thanks for your reply Nicola, I am working in a proof-of-concept before a proper "patch" can be shared.
After some testing I will come back to you guys.

Best regards,

Arturo González

--
You received this message because you are subscribed to the Google Groups "ns-3-users" group.

To view this discussion on the web visit https://groups.google.com/d/msg/ns-3-users/-/GYPUe8wy2qoJ.

To post to this group, send email to ns-3-...@googlegroups.com.
To unsubscribe from this group, send email to ns-3-users+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/ns-3-users?hl=en.