PQ-V2X: A Novel Post-Quantum Cryptographic Dataset for Secure Vehicular Communications

Quantum computing threatens the cryptographic foundations of Vehicle-to-Everything (V2X) security. We present PQ-V2X, a large-scale dataset and benchmarking suite for post-quantum intrusion detection that integrates cryptographic artifacts, vehicular mobility, and zone-aware channel conditions. The dataset comprises five million labeled records spanning 23 attack types across post-quantum, hybrid, and classical families, including 10.05% real traces generated from Kyber, Dilithium, Falcon, and SPHINCS+. We evaluate ten machine-learning models under synthetic and hybrid settings and introduce the Detection Efficiency–Resource (DER) metric to capture accuracy–cost trade-offs. LightGBM attains perfect detection with high DER while training substantially faster than CatBoost; non-boosted models close the classical–quantum gap as data scales from 100K to 200K, though SPHINCS+-based vectors remain challenging. Tunnel zones amplify risk, underscoring the value of zone-aware features. PQ-V2X is, to our knowledge, the first V2X dataset to combine NIST PQC traces, hybrid threats, mobility, and resource-aware benchmarking; data and code are released to enable reproducible, quantum-resilient IDS research.