Intrusion Detection in Wireless Mesh Networks Using a Hybrid Approach

Wireless mesh network are very vulnerable to attacks by malicious users due to, mainly, the cooperation among their nodes. The methods for detection of attacks are grouped into two categories: analysis of anomalies and signatures. The former detects possible new attacks, which have not been previously discovered and the latter depends on prior knowledge to classify the attacks. This paper presents a proposal that uses wavelets and neural networks for detection and classification of attacks. Evaluations by simulation and environmental testing are also presented, and the results indicate that this approach is very promising.