ABSTRACT Vehicular networks are central to intelligent transportation systems, yet their expanding attack surface yields complex multivulnerability paths that exceed traditional detection capabilities. This article presents a high‐coverage attack‐chain generation framework that models, clusters, and composes vulnerabilities with reduced redundancy. A five‐tuple schema, Conditions, Tech, Tool, Target, and Results, unifies heterogeneous sources to expose underlying correlations. A semantically enhanced K‐Modes method integrates semantic similarity with Hamming distance to improve categorization. A two‐dimensional combinatorial engine then generates attack paths using a coverage matrix and greedy selection to balance tractability and coverage. Experiments on communication, system, and control‐layer vulnerabilities achieve over 70% coverage and 85% accuracy, outperforming baselines. The framework supports vulnerability assessment and proactive defense planning, strengthening vehicular network resilience against complex attack chains.
Fang et al. (Fri,) studied this question.