2

I have doubts about the generation of collisions in Veins. I am running the example scenario with only sending beacons, without RSU and without shadowing. I am repeating the simulations varying the beacon interval (from 1 beacon per second (1 second) to 30 per second (0.033 second)) and transmission power (from 5 mW to 40 mW) to see the effect in collisions. What I expected was get few collisions with less txPower and get more collisions if the txPower value is greater. But I got some diffent values for collisions, for example this is the output from some of the simulatios I made:

[*] beaconInterval: 0.5 txPower: 5     Collisions: 0    Total Losts: 0  Received: 114723
[*] beaconInterval: 0.5 txPower: 6     Collisions: 0    Total Losts: 0  Received: 128635
[*] beaconInterval: 0.5 txPower: 7     Collisions: 3    Total Losts: 3  Received: 135691
[*] beaconInterval: 0.5 txPower: 8     Collisions: 0    Total Losts: 0  Received: 142921
[*] beaconInterval: 0.5 txPower: 9     Collisions: 0    Total Losts: 0  Received: 153000
[*] beaconInterval: 0.5 txPower: 10     Collisions: 0   Total Losts: 0  Received: 158094
[*] beaconInterval: 0.5 txPower: 11     Collisions: 0   Total Losts: 0  Received: 166895
[*] beaconInterval: 0.5 txPower: 12     Collisions: 0   Total Losts: 0  Received: 172933
[*] beaconInterval: 0.5 txPower: 13     Collisions: 13  Total Losts: 13 Received: 178446
[*] beaconInterval: 0.5 txPower: 14     Collisions: 745 Total Losts: 1683   Received: 182008
[*] beaconInterval: 0.5 txPower: 15     Collisions: 0   Total Losts: 0  Received: 192333
[*] beaconInterval: 0.5 txPower: 16     Collisions: 0   Total Losts: 0  Received: 196579
[*] beaconInterval: 0.5 txPower: 17     Collisions: 0   Total Losts: 0  Received: 203119

Is it normal to get (for example) 745 collisions with 2 beacons ṕer secod and 14 mW of txPower, and get 0 collisions with the same interval and greater txPower? What is the explanation of these results?

[ADITIONAL INFO: Code]

  • This is the code of veins (4.7.1), I just omit the RSU and shadowing in the scenario and get statics for collisions. The application is based on MyVeinsApp, it just prints when a beacon is received and nothing more.
  • This is the Python code to run simulations varying beaconInterval and txPower (run the script sensibilityOfChanges.py).

In the file resultados.txt you can see the partial results i got. In this file and in the code I refer to Ftx as beacon rate (beaconInterval is obtained by 1/Ftx) and Ptx as txPower.

Francisco Aguirre
  • 83
  • 6
  • Are you able to share a minimal working example (maybe on Github) that would allow someone to reproduce your results? I was not able to. – Christoph Sommer Sep 05 '18 at 15:48
  • Hi @crhistoph, I have added the code I am using in the question. Thanks for your help. – Francisco Aguirre Sep 05 '18 at 20:57
  • I'm currently travelling, so cannot try, but maybe changing the pseudorandom seed (or just averaging over multiple independently seeded repetitions) might shed some light on this. – Christoph Sommer Sep 06 '18 at 09:34
  • On a completely unrelated note, you might want to look into the OMNeT++ way of doing parameter studies. This allows you to use one .ini file for setting up all simulations (varying both parameters and seeds automatically) – Christoph Sommer Sep 06 '18 at 09:35
  • So, it is not possible to find the max txPower without collisions due to the random generation of collisions? in other words, it is possible to get 0 collisions in a scenario with vehicles transmitting with 10mW, get 10 collisions in the same scenario with vehicles at txPower = 11mW and get 0 collisions with txPower at 12mW? I was thinking to make some binary search before seeing this behavior, something like this https://www.tamps.cinvestav.mx/~faguirre/vanets/colis.png – Francisco Aguirre Sep 06 '18 at 18:47

0 Answers0