An ARIMA-Based Approach for Predicting and Detecting Routing Attacks in MANETs

Abstract

Mobile Ad-hoc Networks (MANETs) are wireless networks in which a group of mobile devices, or nodes, communicate without relying on a centralized infrastructure. MANETs have become particularly important in scenarios where traditional network deployments are unavailable, impractical, or costly. One of the most critical challenges in MANETs today is ensuring secure communication, as these networks are vulnerable to attacks at the physical, network, and application layers. Our research addresses security issues in MANET routing protocols by developing innovative techniques for detecting and predicting routing attacks. This paper proposes a novel approach using the Auto-Regressive Integrated Moving Average (ARIMA) model, a well-established statistical method for time series analysis and forecasting. The proposed model aims to predict routing attacks in MANETs, thereby enhancing the security and reliability of these highly dynamic wireless networks. Our model is evaluated using various performance metrics, including accuracy, specificity, precision, and sensitivity, all of which achieve an average score of 0.98. Furthermore, a comparative analysis with the Artificial Neural Network (ANN) model in the same attack environment, as well as with other existing methods, demonstrates the superiority and effectiveness of the proposed ARIMA model in detecting and predicting routing attacks in MANETs.