Hierarchical Architecture for Computational Offloading in Autonomous Vehicle Environment

Abstract

Mobile Edge Computing (MEC) is a key enabler technology for fifth generation (5G) networks and has numerous use cases including, Device-to-Device (D2D) communications and computation offloading. In the near future, Internet of Vehicles (IoV) applications will require high data rate as well as extensive computational resources. In the connected vehicles, MEC has emerged as a strong candidate due to its proximity with the users, high throughput, better traffic monitoring & management, large coverage area, and context-awareness. For this purpose, a vehicular architecture requires to handle the computation under stringent latency conditions and meet high computational requirements. This paper proposes a hierarchical architecture for computation offloading for future vehicular network. The proposed architecture divides the computation offloading into multiple levels, resulting in efficient and cost-effective architecture. Furthermore, we propose to make decision for each task based on speed, computational requirement and latency. We assume that a controller as an application is installed within the MEC server to handle the computation handover efficiently without introducing complexity into the network.