Tactical Manet Attack Detection Based on Fuzzy Sets Using Agent Communication

摘要

This study describes a distributed attack detection approach for a tactical MANET using intelligent agents equipped with inference systems based on fuzzy logic. The results produce a prototype intrusion detection system capable of effectively detecting attacks in a tactical MANET with accuracy approaching 95%. The attack recognition system is implemented using stationary intelligent fuzzy agents (SIFA) resident on each node. Agents run autonomously on each node, collect packets from the data stream, extract relevant information, exchange information through light-weight messages, and trigger alerts using the fuzzy inference process.