Transport anomalies simulation and analysis in wireless networks

Abstract

In the paper we research the affordable simulation system for gathering network traffic and analyzing its anomalies for TCP-based wireless networks. We consider networks that include enough number of robots, drones and other unmanned aerial vehicles (UAVs), and which need to be analyzed for traffic transport prob lems due to hardware/software failures or external impacts. The simulation system should be able to gather traffic information necessary for further analysis by machine learning techniques and providing feedback for network operators. We considered to use and extend GrADyS-SIM package for OMNeT++ simulation envi ronment, as it was specially targeted for UAVs networks. The simulated networks traffic is preprocessed for extracting valuable information and adding timing marks for further analysis in unsupervised learning-based classifier built using the OPTICS methods. The rich tools, integrated into OMNet++ system provide simplified traffic scenarios creation and good visual presentation. The classifier provides traffic flow events clustering and anomalies detection. We made series of experiments, proving that the classification results are valuable for network operation and further may be used for automatically tuning traffic volume from UAVs to operators and servers.

Problems in programming 2026; 3: 20-28