This chapter reviews the latest progress in the design of edge computing systems for autonomous driving applications. It begins with benchmark suites available for evaluating edge computing system designs. The chapter then reviews different approaches in designing computer architectures for autonomous driving workloads. Next, it describes the designs of runtime layers for efficient mapping of incoming workloads onto heterogeneous computing units. The chapter also discusses the designs of middleware for binding different autonomous driving functional modules. The middleware layer facilitates the communication between different autonomous driving services. The chapter summarizes several challenges. First, the middleware should impose minimal computing overhead and memory footprint, thus making it scalable. Secondly, as some autonomous driving services may stay in edge clouds, the middleware should enable a smooth edge client and cloud communication. The chapter further presents real‐world implementations of autonomous driving edge computing systems.
Edge Computing for Autonomous Vehicles
2020-04-13
11 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
DYNAMIC EDGE COMPUTING WITH RESOURCE ALLOCATION TARGETING AUTONOMOUS VEHICLES
Europäisches Patentamt | 2023
|Secure Multi-access Edge Computing Assisted Maneuver Control for Autonomous Vehicles
BASE | 2021
|