A jamming detection method for multi-hop wireless networks based on association graph

Abstract

In this paper, a jamming detection algorithm based on association graph is put forward based on the observation that different jamming attacks will cause different network status changes in multi-hop wireless networks (MHWNs). The proposed algorithm consists of two phases, i.e., learning and detection phases. At the learning phase, different symptoms are extracted through learning from various samples collected in both jamming and jamming-free scenarios. Then, a symptom-attack association graph is built. At the detection phase, the association graph is adopted to detect the jamming attacks that lead to the observed symptoms. A series of simulation experiments on NS3 have validated that the proposed method can efficiently detect and classify typical jamming attacks, including reactive, random and constant jamming attacks.