Artificial Noise Assisted Secure Mobile Crowd Computing in Intelligently Connected Vehicular Networks

Growing computation requirements in emerging intelligently connected vehicle (ICV) networks has posed a big challenge for vehicles to process computation-intensive tasks on-board, and thus computation partition and offloading plays a critical role. Owing to broadcast nature of wireless channels, confidential messages without proper security protection in ICVs may be easily exposed to malicious users. In this paper, a mobile crowdsourcing (MCS) based mobile crowd computing framework for ICV networks is proposed, where multiple vehicles act as workers to provide computing services for end user (EU). In particular, artificial noise (AN) assisted physical (PHY) layer security approaches are used to enhance the security in offloading links. Ergodic secrecy rates in different offloading phases in time-varying channels are derived. In addition, an incentive-driven mechanism is introduced to encourage workers to share their idle computing resources, and an optimization problem is formulated to minimize the overall price paid for computing tasks, subject to energy consumption and delay constraints. Finally, Monte Carlo simulations verify the analysis on the ergodic secrecy rates.