RLoRa

This repo is my working folder for a LoRa MAC protocol study in OMNeT++. I keep the simulation model in src/, configs in simulations/, and scripts to export/aggregate data in scripts/.

Below is a guide for running the whole pipeline, especially the OMNeT++ simulation campaign.

Repo layout (quick map)

• src/: C++/NED model code for the LoRa network and MAC protocols.
• simulations/: OMNeT++ configs (omnetpp.ini + parameter files).
• scripts/shell/: run scripts for the campaign + export helpers.
• scripts/python/: Python aggregation scripts for the exported JSON data.
• data/: exported per-run JSON vectors (input for aggregation).
• data_aggregated/: aggregated metrics (what I plot).
• out/: build output (binary ends up in out/.../src/rlora).

Prerequisites

• OMNeT++ (for opp_runall, opp_run, opp_scavetool).
• INET framework (this repo expects it at ../inet4.4 by default).
• C++ build toolchain for OMNeT++.

The Makefile uses INET_ROOT, so go into INET's directory and run source setenv. Do the same for Omnet++.

Some scripts also use rlora_root:

export rlora_root=/path/to/rlora

Build the simulation binary

From the repo root:

make cleanall
make makefiles
make

This produces the runnable binary under out/<mode>/src/rlora.
The scripts here reference out/clang-release/src/rlora, so adjust if needed.

OMNeT++ simulation campaign (main part)

The campaign is defined in:

• simulations/omnetpp.ini
• simulations/staticParams.ini
• simulations/dynamicParams.ini
• simulations/statistics.ini

Key points:

• Parameter sweeps for numberNodes and ttnm (time-to-next-mission).
• Area sizes: 300m, 1000m, 5000m, 10000m.
• MAC protocols: Aloha, Csma, MeshRouter, IRSMiTra, RSMiTra, RSMiTraNR, MiRS, RSMiTraNAV.
• Mobility configs: MassMobility and GaussMarkovMobility.
• Output vectors/scalars are written to simulations/results/ by default.

Run a small test (single run)

This is how I sanity-check:

opp_run -u Cmdenv -n ./simulations:./src:./../inet4.4/src \
  -l ./../inet4.4/src/INET \
  -f ./simulations/omnetpp.ini \
  -c MassMobility \
  -r 0

Adjust the INET path if it is not at ../inet4.4.

Run the full campaign

I use the opp_runall scripts in scripts/shell/:

./scripts/shell/start-1.sh   # MassMobility
./scripts/shell/start-2.sh   # GaussMarkovMobility

Notes:

• They use -j80 (80 parallel jobs). Reduce if your machine is smaller.
• They run a big range of run numbers (-r 243200..255999) that cover the parameter sweep. This is heavy and takes a long time.
• If you want to regenerate the start scripts, use:

./scripts/shell/populate-start-script.sh

Export vectors/scalars to JSON

After simulations finish, the results are in simulations/results/ as .vec and .txt. Export them to JSON with:

./scripts/shell/exportData.sh

This script expects rlora_root to be set and writes JSON files into data/ under protocol/dimension folders (e.g., data/Aloha/300m/...).

Aggregate metrics (Python)

Run the Python aggregators from the repo root. Example for node reachability:

python scripts/python/data-evaluation/aggregate_metrics/aggregate_node_reachability.py

Other aggregators live in:

scripts/python/data-evaluation/aggregate_metrics/

Outputs land in data_aggregated/, grouped by metric.

Typical workflow

1. Build (make cleanall, make makefiles, make).
2. Run campaign (start-1.sh, start-2.sh).
3. Export with exportData.sh.
4. Aggregate metrics with the Python scripts.
5. Plot or inspect the JSON in data_aggregated/.

If anything fails, the first things to check are:

• Is INET_ROOT pointing to the right INET version?
• Is rlora_root set?
• Are the INET paths in the start scripts correct for the machine?