Context-Intelligent Predictive Handover and Ephemeral Rekeying for Secure Sensor-Assisted V2X Communication

Emergent autonomous vehicle networks rely on persistent sensor-based V2X connectivity. However, high-speed mobility, along with handovers and repetitive key-establishment leads to authentication latency, packet-loss, and security vulnerabilities. In this paper, we introduce CIPHER-Drive, A Context-Intelligent Predictive Handover and Ephemeral Rekeying framework which facilitates seamless secured connectivity i.e., achieving a secured state before completion of vehicular handover from a Road Side Unit (RSU). To accomplish this feat, mobility-context information is abstracted from vehicular velocity, direction of travel, received signal strength, traffic-load, link quality, and vehicular sensor measurements. This information is further used to predictively derive the future RSU of attachment and the appropriate level of security. Then security specific services such as predictive pre-authentication and context-dependent ephemeral session-keying are executed. Cipher-drive prediction model is compared with QELP and SACO-PRESENT techniques under varying V2X conditions utilizing standard communication and security protocols. Cipher-drive shows a 94.8% probability of successfully completing secure handovers, 94.3% packet delivery ratio, 93.6% key-establishment ratio, 92.9% reduction in signaling-overhead and 94.5% accuracy in handover prediction. The results indicate reduced security latency with minimal compromise to reliable packet delivery by using predictive authentication and ephemeral context-bound session keys. We show that Cipher-drive is able to provide a lightweight, mobility-aware, and context-adaptive security service for sensor-enabled autonomous vehicular networks. Predictive model further helps to minimize unnecessary executions of complete handshakes allowing for increased session freshness and packet delivery in high density traffic scenarios, RSU handovers, and under attack.