FANET Jamming Detection ML Framework

Machine Learning Framework & Automated Dataset Generation for FANET Jamming Detection (NS-3)

Hello,

I am currently working on a research project titled:

“Intelligent Jamming Detection in Flying Ad Hoc Networks (FANETs) Using NS-3 Simulation and Machine Learning.”

I have already completed the NS-3 simulation phase and now need support with the machine learning, automated dataset generation, and data processing phase.

⸻

What Has Already Been Completed

* Built FANET scenarios in NS-3
* Implemented the Gauss-Markov mobility model
* Implemented multiple jamming attack types:
* Reactive jammer
* Hybrid jammer
* Constant jammer
* Random jammer
* Executed simulations under multiple network conditions
* Generated network performance metrics from NS-3 outputs

⸻

Current Simulation Outputs

The simulation outputs currently include metrics such as:

* Total transmitted packets
* Total received packets
* End-to-end delay
* Packet loss
* Packet Delivery Ratio (PDR)
* Packet Loss Ratio (PLR)
* Throughput
* RSSI / RSSI in dBm
* Possibly SINR and additional metrics later

I also have simulation scenarios with the detection algorithm enabled and disabled.

⸻

Example Metrics Per Scenario

Hybrid Jammer

* Tx packets
* Rx packets
* Delay
* Throughput
* RSSI
* PDR
* PLR

Reactive Jammer

* Same metrics as above

⸻

Main Objective

I need to transform the current NS-3 simulation outputs into a complete machine learning framework for intelligent jamming detection in FANETs.

⸻

What I Need

1. Automated Dataset Generation from NS-3

I currently have single/manual simulation runs working successfully. However, I now need to scale the framework to automatically generate large machine learning datasets (thousands of labeled samples) directly from NS-3 simulation outputs.

Required Features

* Automatically execute multiple simulation scenarios
* Automatically vary simulation parameters between runs
* Automatically extract metrics from NS-3 outputs
* Automatically generate CSV datasets
* Automatically assign labels for machine learning

Parameters That May Change Automatically

Examples include:

* UAV/node speed
* Number of UAVs
* Simulation duration
* Jammer type:
* Reactive
* Hybrid
* Constant
* Random
* Jammer power
* Mobility conditions
* Traffic rate
* Detection algorithm enabled/disabled
* RSSI/SINR conditions
* Transmission range

Preferred Integration

The automation should preferably be integrated directly with:

* input.cc
* NS-3 simulation scripts
* Output trace files/log files

The goal is to avoid manually running and labeling simulations one by one.

⸻

2. Dataset Preparation

Transform and organize all NS-3 outputs into structured machine learning datasets (CSV format).

Example Dataset Columns

* TxPackets
* RxPackets
* DelayMs
* LostPackets
* PDR
* PLR
* ThroughputKbps
* RSSI_dBm
* DetectionAlgorithm
* JammerType
* Label

Example Labels

* Normal
* Reactive_Jamming
* Hybrid_Jamming
* Constant_Jamming
* Random_Jamming

⸻

3. Data Preprocessing

Including:

* Data cleaning
* Handling missing values
* Feature normalization/scaling
* Label encoding
* Feature selection (if needed)
* Train/test split

⸻

4. Machine Learning Implementation

Implement and compare ML models for jamming detection, including:

* Random Forest
* SVM
* k-NN

(Optional later)

* LSTM
* CNN
* Deep Learning models using TensorFlow/Keras

⸻

5. Model Evaluation

Evaluate the models using:

* Accuracy
* Precision
* Recall
* F1-score
* Confusion matrix
* Detection latency (if possible)

⸻

6. Deliverables

Please provide:

* Python source code
* Well-commented scripts
* Automated dataset generation scripts
* CSV dataset generation pipeline
* Documentation/explanations
* Graphs and visualizations
* Model comparison results

⸻

Preferred Tools/Libraries

* Python
* Pandas
* Scikit-learn
* Matplotlib
* TensorFlow/Keras
* Jupyter Notebook

⸻

Important Notes

* I have already completed the NS-3 simulation development phase.
* I do NOT need help building FANET simulations from scratch.
* The main requirement is automating dataset generation from NS-3 outputs and integrating the data into a machine learning framework.
* Experience with NS-3, wireless networks, FANETs, cybersecurity, or network intrusion/jamming detection is highly preferred.

⸻

Please Include in Your Proposal

* What information/files you need from me
* Estimated timeline
* Estimated cost
* Your experience with:
* NS-3
* Machine Learning
* Network Security
* FANETs
* Wireless Network Datasets
* Automated simulation/data pipelines

Thank you.