AI-Empowered Content Caching in Vehicular Edge Computing: Opportunities and Challenges
Muhammad Awais Javed, Sherali Zeadally
Abstract
Vehicular networks are an indispensable component of future autonomous and intelligent transport systems. Today, many vehicular networking applications are emerging, and therefore, efficient data computation, storage, and retrieval solutions are needed. Vehicular edge computing (VEC) is a promising technique that uses roadside units to act as edge servers for caching and task offloading purposes. We present a task-based architecture of content caching in VEC, where three major tasks are identified, namely, content popularity prediction, content placement in the cache, and content retrieval from the cache. We present an overview of how artificial intelligence techniques such as regression and deep Q-learning can improve the efficiency of these tasks. We also highlight related future research opportunities in areas such as collaborative data sharing for improved caching, efficient sub-channel allocation for content retrieval in C-V2X, and secure caching.