A Novel Approach for Estimating Performance of IIoT-Based Virtual Control Train Sets under DoS Attacks

Abstract

The virtually coupled train sets (VCTS) have been proposed to improve operational capability and passenger satisfaction and ensure punctuality, thus alleviating the rapidly worsening traffic pressure. Recently, due to the lack of reliable wireless communications and accurate perceptual information, VCTS based on industrial internet of things (IIoT) are receiving growing concerns by integrating into the IIoT, AI, and edge computing. However, denial of service (DoS) attacks are feasible for IIoT-based VCTS due to the physically exposed open electromagnetic environment. They would cause severe safety and punctuality problems, such as poor real-time capability, enormous packet dropout rates, extensive train operational delays, and disturbance in the train convoy’s dynamic schedule. This paper deeply discusses the effects of DoS attacks on the performances of the IIoT-based VCTS by combing the physical layer with the cyber layer and explores the requirement of an attacker. We consider that the system is under the attack of a rational DoS attacker with limited jamming attacks, which will cause the most system state offset. In the paper, we propose a novel train status estimation approach to compensate for the losing information of the front train by the trade-off between the best gain of the DoS attacker and the punctuality of the IIoT-based VCTS. System performance includes physical dynamic indicators, train operational delay variance, and average waiting time of passengers. Taken together, these findings indicate that the established status estimation approach can effectively mitigate safety concerns and reduce train operational delays.