Shadowing buildings in a Vanet of a real urban area (Veins, OMNet++, SUMO)

[Simone D'Arpino]

Hi,

I'm trying to understand why two stationary vehicles  of a Vanet in a real urban area don't establish a bidirectional communication in two different instant of the simulation (always remaining stationary) : the node A, in fact, transmit a message to node B that receive correctly the message, but when B transmit its packet, A doesn't receive it. This problem is due to path loss model that i use to take into account the presence of buildings, or it's due to some fading model that is implemented in Veins?

Thanks in advance to all those who want to give their help.

Simone

[Christoph Sommer]

Simone D'Arpino wrote:
> node A, in fact, transmit a message to node B that receive correctly the
> message, but when B transmit its packet, A doesn't receive it

Are these the only two nodes in your simulation?

If so, maybe different transmit powers are to blame.

If not, maybe another node is interfering with the transmission.


Best,

Christoph

--
Dr.-Ing. Christoph Sommer
Institute of Computer Science
University of Innsbruck, Austria
Phone: +43 512 507-96835 / Fax: -9888
http://ccs.uibk.ac.at/~sommer/

[Simone D'Arpino]

Hi Dr.Sommer,

I'm sorry for the superficiality of may previous post, now I will try to be clearer: the real urban area considered is a district of Rome (2km x 2km), then there are many buildings that disturb the inter-vehicular communications and there are more of 600 vehicles (all stationary) which in turn, one at a time, transmit a broadcast packet. I used either Simple Path Loss Model and Two Ray Interference Model with Simple Obstacle Shadowing, the transmission power is 500 mW and the bit-rate is 6 Mbps. Unfortunately, there are more than one hundred vehicles that don't receive packets from vehicles that instead had previously received the messages transmitted by those vehicles. One possible explanation is that the SNIR of the communication between two vehicles exceeds the threshold for the processing of the message in only one case but not the other, but I'm not so sure.

Thank you so much,

Simone

[Christoph Sommer]

Simone D'Arpino wrote:
> 600 vehicles [...], one at a time, transmit a broadcast packet.

Depending on how you configured your simulation, this may be an
illusion. In the GUI you will always see only one event being processed
at a time. You will need to closely inspect the beginning and end times
of frames if you want to be sure that you always receive the event
corresponding to the end of one frame before that of the beginning of
the next (begin/end/begin/end instead of begin/begin/end/end, which
would indicate overlapping frames).

> One possible explanation is that the SNIR of the communication
> between two vehicles exceeds the threshold for the processing of the
> message in only one case but not the other, but I'm not so sure.

You can investigate by looking at the debug output (if you run the
simulation with debug output turned on) or by inspecting with a
debugger: if the signal level of a received transmission is low, a
building is to blame for the packet loss. If the signal is high, but the
SINR is low, an interfering transmission is to blame.

Best,

Christoph

--
Dr.-Ing. Christoph Sommer
Institute of Computer Science
University of Innsbruck, Austria

Phone: +43 512 507-53285 / Fax: -53079
http://ccs.uibk.ac.at/~sommer/

[Simone D'Arpino]

Hi Dr.Sommer,

first of all, thank you soo much for the attention and the help. I launched my simulation again and my scalar statistic are very clear: every vehicle transmit its packet wihtout contention with other vehicles, in fact, scalar statistic SentPacket is one for each of them; every vehicles have the scalar statistic SNIRLostPacket greater than zero, then there some discarded packets because the level of received signals aren't so much larger than that of noise and interferences, le statistic scalar ncollisions, DroppedPacketInMAC and numInternalContention are all equal to zero, then every vehicle doesn't receive more than one packet at a time and transmit correctly its message, and finally, the scalar statistic TotalBusyTime is lower that the interval between transmission of two vehicles. Therefore, I ask: since a packet is received correctly judged only when the signal level exceeds the threshold set as SNIR, in Simple Path Loss Model or Two Ray Interference Model and Simple Shadowing Buildings Model there is a there is an element of uncertainty that could cause such results? Because if so, I definitely consider reliable my results and go on with my research.

Thanks for the help

Simone D'Arpino

[Christoph Sommer]

Dear Simone,

Simone D'Arpino wrote:
> in Simple Path Loss Model or Two Ray Interference Model and Simple
> Shadowing Buildings
> Model there is a there is an element of uncertainty that could cause
> such
> results?

I do not think that any of these path loss models has stochastic
properties. They should be fully deterministic.

However, the default Decider component [1] in Veins is stochastic. See,
for example, [2].

Best,

Christoph




[1]
https://github.com/sommer/veins/blob/veins-2.2/examples/veins/config.xml
[2]
https://github.com/sommer/veins/blob/veins-2.2/src/modules/phy/Decider80211p.cc#L334

[jack ouimet]

Dear Simon, I was looking to understand how to modelete a real urban area. My research aim is to compute the attenuation signal of received 

transmissions in an urban area, means presence of buildings that cause interference. 

However, i get confused which analog model to use or  which obstacle model to use??

This is my actual simulation parameters. I have this analog model configuration:

<?xml version="1.0" encoding="UTF-8"?>

<root>

<AnalogueModel type="LogNormalShadowing">

<!-- Mean attenuation in dB -->

<parameter name="mean" type="double" value="0.5"/>

<!-- Standart deviation of the attenuation in dB -->

<parameter name="stdDev" type="double" value="0.25"/>

<!-- Interval in which to define attenuation for in seconds -->

<parameter name="interval" type="double" value="0.001"/>

</AnalogueModel>

<Decider type="Decider80211p">

<!-- The center frequency on which the phy listens-->

<parameter name="centerFrequency" type="double" value="5.890e9"/>

</Decider>

</root>

However, as i read in veins documentation [1], the simple obstacle model is used to modelete real urban buildings.

 In addition, in your meessage,  you mentioned the Simple Path Loss Model orTwo Ray Interference Model and Simple Shadowing Buildings Model??? 

I want to know if all these models can modelate an urban area (buildings) or you have to add obstacles??
 Really i didn't understand which analog model to use/ obsacle model  in order to modelate real urban area.!! 

It will be a big help for me if i can receive some explanation from you.

[1] http://veins.car2x.org/documentation/modules/obstacles/ 

thanks

jack/perl

--
You received this message because you are subscribed to the Google Groups "omnetpp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to omnetpp+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[loon yun]

Hi there, are you able to simulate packets drops & routes in veins, sumo, and omnet++? Please help. 

Student doing traffic simulation