Nakagami problem

[Gabriele maggio]

hi guys,
I have to realize a comparison between simulations running under ns-3
and simulations running under ns-2 using Nakagami propagation loss
model. Looking at the code of both simulators, ns-3
NakagamiPropagationLossModel is different from ns-2 implementation and
probably for this reason I obtain different results. How can I set
propagation loss model in ns-3 so that it is equal to ns-2 Nakagami
loss model? Do I have to create a chain with
ThreeLogDistancePropagationLossModel?

Thanks, Gabriele

[phuchong kheawchaoom]

Yes, you have to create a chain with ThreeLogDistancePropagationLossModel or LogDistancePropagationLossModel.

Because,  NakagamiPropagationLossModel is only fast fading which does not include distance loss for slow fading in ns-3.

Regards

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

[Gabriele maggio]

OK perfect.. but I have another question: from ns2 Nakagami
propagation loss model I can obtain ns3 Nakagami attribute values, but
what about LogDistance (or ThreeLogDistance) attributes? Do I have to
use default values or can I obtain them from ns-2 Nakagami?

Thanks
Gabriele

[phuchong kheawchaoom]

Honestly, I do not know. You have to look the detail in both ns2 and ns3 then change attribute in ns3 following ns2.

Regards

[Ali R]

Hi Gabriele,

Since it is supposed to return the RxPower, I assume they've forgotten to include the Three Log Distance Path Loss part.

You can download the patch here:

http://winlab.rutgers.edu/~rostami/patches/NakagamiModelPatch.html

-Ali

[Konstantinos]

Dear both,

There is no problem with the Nakagami nor it requires any patch.

The Nakagami-m only gives you the FastFading effect which simply vary the RxPower according to a Nakagami distribution.

In order to have also the effect of distance, then you add the Three Log Distance or any other model you want.

NS-3 provides the API for that in the ChannelHelper class.

https://www.nsnam.org/doxygen/classns3_1_1_yans_wifi_channel_helper.html#a71c64d54e148b43698c65736e1c77156

And with an example:

YansWifiChannelHelper wifiChannel;
wifiChannel.AddPropagationLoss("ns3::LogDistancePropagationLossModel");
wifiChannel.AddPropagationLoss("ns3::NakagamiPropagationLossModel");

and you can add any other parameters/attributes to each of the models.

[Tommaso Pecorella]

I guess that the "bug" is in the documentation. It should be clear from the docs what's a fast fading (i.e., zero mean) and a distance-based model.

I'd open a specific enhancement request for this point, as it's not clear at all from the docs.

T.