Dynamic Weight Allocation–Based Network Security and Anomaly Detection Model for Intelligent VANETs

Abstract

Determining the weights of evaluation metrics is one of the key factors influencing the cybersecurity and anomaly detection of intelligent vehicular ad hoc networks (VANETs). To address the limitations of traditional weighting methods, which often overlook the impact of changes in metric attribute states on evaluation weights, this paper proposes a dynamic weight allocation–based network security and anomaly detection model. The model begins by decomposing and analyzing the security and anomaly detection objectives of VANETs, constructing a comprehensive evaluation metric system. The network security assessment model for VANETs presented in this research overcomes the drawbacks of conventional static models by utilizing a dynamic weight allocation technique. Based on current network conditions, a state variable weight method was created that dynamically computes security values by combining incentive and penalty mechanisms. A ranking-based weighting algorithm is employed to analyze the correlation between security and anomaly detection metrics. Subsequently, the proposed dynamic weight allocation algorithm calculates the dynamic weights of individual metrics within the system, enabling a robust assessment of network security and anomaly detection for intelligent VANETs. The evaluation results provide security level classifications and identify anomalies effectively. Experimental results demonstrate that the model significantly enhances the rationality and accuracy of intelligent VANET evaluations, contributing to improved cybersecurity and anomaly detection.