Establish positions/mobility of UEs and eNBs

[khaledhassan]

Relevant lines of lena-x2-handover-measures.cc: 231-248

• [x] Implement correct mobility model for UEs (Random Walk)
• [x] Put eNBs at a position that makes sense (see eNB_positioning branch) --got hexagonal working, might be easier/make more sense to use simple grid
• [x] eNB grid positioning
  • [x] Replace GridPositionAllocator with a for loop so we can move the eNBs up by ~30m~ 15m or so
  • [x] decide if want to use hex grid
• [x] Install handover algorithm (using rsrp, see handoverAlgorithm branch)

[khaledhassan]

There is a function

void LteHelper::AttachToClosestEnb (NetDeviceContainer ueDevices, NetDeviceContainer enbDevices)

in lte-helper.cc:1018 that could help us automatically attach the right ue to the right enb

[khaledhassan]

Handover algorithm code:

  lteHelper->SetHandoverAlgorithmType ("ns3::A3RsrpHandoverAlgorithm");
  lteHelper->SetHandoverAlgorithmAttribute ("Hysteresis",
                                            DoubleValue (3.0));
  lteHelper->SetHandoverAlgorithmAttribute ("TimeToTrigger",
                                            TimeValue (MilliSeconds (256)));

Put this after:

  Config::SetDefault ("ns3::LteEnbPhy::TxPower", DoubleValue (enbTxPowerDbm));
  enbLteDevs = lteHelper->InstallEnbDevice (enbNodes);
  ueLteDevs = lteHelper->InstallUeDevice (ueNodes);

[khaledhassan]

Hex grid for eNBs: see file lte-hex-grid-enb-topology-helper.cc

[slichtenheld]

Waypoint model may be better to trigger handovers, don't think random walk is the best model here.

[khaledhassan]

That's true, but the paper used the random walk model. Perhaps we can do a comparison between them as a part of our report. I think/hope we can get the random walk working with real values for the eNB distances, such as the ones they used in the paper