Secure Digital Twin Migration in Edge-Based Autonomous Driving System

Abstract

Digital twin technology is being applied increasingly in the Internet of Vehicles environment, but it still faces many challenges in terms of efficiency and security. In the field of digital twin-based autonomous driving, many previous works have been done to study efficient migration methods of digital twin models. But these works consider the migration process as a blackbox. We study the efficient migration method of the digital twin model between the edge computing nodes inside the blackbox. We propose three different migration strategies depending on the source of the initial data and the source of the updated data, and evaluate the efficiency of these strategies in terms of migration time in different network environments using the autonomous driving simulation platform CARLA. We then derive methods for selecting migration strategies under different network conditions. During the migration process, there may be external attacks on participating elements or networks. We analyze the security problems that may arise during the migration process and propose corresponding defense methods against such cyber attacks.