MobilityHelper

[Sarah Dumoulin]

Hi,

  I'm very new to NS3, and I'm looking at an existing model and trying to implement some changes.  However, I'm a little puzzled by the behaviour being exhibited and I was hoping for some pointers on how to approach the problem.

  The existing model set (Wifi / LTE coexistence development branch) has a simple example network made of an LTE node and a UE and a WIFI node and a UE, with two flows sent during the simulation (one between each pair).  When I run the unmodified code I get a certain number of packets received at the UE in each pair.

  However, I want to modify this model (to have an adaptive duty cycle), so I started to change some files.  Once I introduce a new MobilityHelper in the scenario-helper.cc file the number of packets received by each UE during the simulation is halved.  I don't have to do anything with the helper.  I just add the line "MobilityHelper mobility;" to this file in the "ConfigureLte" function and the behaviour changes.  I don't have to install the helper in order to trigger the change in behaviour.  If I enable logging for the MobilityHelper module I see that the new MobilityHelper I created is not installed.  If I comment out the MobilityHelper line, the behaviour of the model reverts back to normal.  (I've tried using a different variable name.  I also tried installing the mobilityHelper with the ns3::ConstantPositionMobilityModel, as there is no mobility in this simulation.  Neither had any effect).

  Obviously, I need the mobilityHelper in order to make other changes.  I wish to install nodes to monitor the DL channel at the enBs.  This worked under the previous LAA branch of the Wifi/LTE coexistence dev branch.  However, I'm stumped as to why the mere existence of a MobilityHelper within this application would cause such a change in behaviour in the final simulation.  Can anybody offer me any insight?

Thanks,

Sarah

[Tommaso Pecorella]

Hi,

the mobility helper could have changed the position of the nodes in a way you're not expecting. Consider that including the MobilityHelper will (most probably) make the nodes moving... and this is going to have some results.

Of course, this is true if the nodes are moving. If you're using a ConstantPositionMobilityModel... then there's something else.

Hope this helps,

T.

[Sarah Dumoulin]

Hi,

  Thanks for your response.  That was also my first thought, which is why I tried assigning the ConstantPositionMobilityModel to the MobilityHelper.  Unfortunately, this made no difference to the simulation results. 

  

  I can try to enable further logging to ensure that nothing is moving around that isn't supposed to be, though.

Thanks again,

Sarah