Grey Model and Polynomial Regression for Identifying Malicious Nodes in MANETs

Abstract

Nodes positioning is an essential issue for diverse applications in Mobile Ad Hoc Networks (MANETs). However, besides misbehaving nodes that could cause power depletion, MANETs are also susceptible to cyber-attacks, which can make the network unstable and/or unavailable. Therefore, considering the gaps aforementioned, the goal of this paper is to propose a model for identifying malicious/misbehaving nodes by: (1) the use of two forecasting methods (Grey Model and Polynomial Regression); (2) variability analysis; and (3) simulation of fake node positions. The obtained results allow concluding our model has high rate of accuracy for detecting malicious/misbehaving nodes.