Fuzzy-Based Trust Model for Detection of Selfish Nodes in MANETs

Abstract

Cooperation among nodes is mandatory for smooth operation of Mobile Ad Hoc Networks (MANETs) in terms of data routing. A participating node may refuse to deplete its resources for the benefit of others because of not getting any direct advantage for its service. Nodes showing such behavior are called selfish or non-cooperative nodes. Non-cooperative nodes can severely affect the performance of MANETs. Non-cooperative behavior of nodes in the MANETs may lead to network partitioning. In this paper, we address the issue of non-cooperative behavior by incorporating the concept of fuzzy logic closely coupled with the concept of trust. Fuzzy-based analyzer is used to distinguish between the non-cooperative behavior nodes and trustworthy nodes. We propose a fuzzy-based scheme to detect and isolate non-cooperative nodes in MANETs. In the proposed scheme, every node in the network constantly monitors its one-hop neighbors for their actions. Every node computes the trust of the observed neighbors. These trust values are passed on to a fuzzy function which is mapped into different classes. The resulting classes show the trust levels of the observed nodes. On the basis of the calculated trust value, the non-cooperative nodes are detected and isolated from the active routes of the MANET. Proposed fuzzy-based scheme is robust enough in terms of detecting packet drop attack in the network. Results show that proposed scheme detects non-cooperative nodes effectively with low false positives rate. Moreover, proposed scheme increases the packet delivery ratio and throughput in the presence of non-cooperative nodes in the network.