SMART: A SVM-based Misbehavior Detection and Trust Management Framework for Mobile Ad hoc Networks

Due to lack of pre-deployed infrastructure, nodes in Mobile Ad hoc Networks (MANETs) are required to relay data packets for other nodes to enable multi-hop communication between nodes that are not in radio range with each other. However, whether for selfish or malicious purposes, a node may refuse to cooperate during the network operations or even attempt to interrupt them, both of which have been recognized as misbehaviors. To address the security threats caused by various misbehaviors, a SVM-based Misbehavior Detection and Trust Management framework (SMART) is studied in this paper. In SMART, Support Vector Machine (SVM) technique is used to detect node misbehaviors, which requires neither a pre-defined threshold nor any well-formed behavioral pattern to distinct misbehaviors from normal behaviors. In addition, multi-dimensional trust management scheme is applied evaluate the trustworthiness of MANET node from multiple perspectives. By this means, the trustworthiness of each node can be described in a more accurate and effective manner. To validate the proposed SMART framework, an extensive performance study is conducted in terms of network simulation. The simulation results show that the SMART framework yields a good performance in various scenarios.