I am currently working on modeling a Sybil attack in a Vehicular network using NetSim. My goal is to detect the attack using a machine learning classifier. I would like guidance on how to accurately model the Sybil attack within NetSim and how to integrate a machine learning-based detection method, using features extracted from the simulated network data.
An ML classifier can use the fact that multiple nodes have approximately the same RSSI at the same time to detect a Sybil attack. This scenario is indicative of a Sybil attack because, under normal circumstances, different vehicles passing the same RSU should have varying RSSI values due to differences in their locations, speeds, and orientations.
In NetSim, first, create multiple malicious (Sybil) nodes that mimic legitimate vehicles. Then, calculate the pairwise RSSI differences between vehicles at each RSU, and assign a threshold to classify vehicle pairs based on these differences. Label the similarity score as 0 or 1, depending on the RSSI similarity. Use the labeled data to train a classifier. Finally, apply the trained classifier to detect Sybil attacks in other scenarios
See the section “AI/ML for Sybil Attack Detection in VANETs” in the page https://tetcos.com/machine-learning-netsim.html
- Badges
- Science topic
- Similar topics
- Biological Science
- Bioecology
- Systems Ecology
- Landscape Ecology
- Connectivity