In-vehicle network latency analysis for a lane keeping assistance system

Abstract

Due to the rapid advancements in automotive technologies, vehicles now rely on additional high-speed sensors. This development has led to an increase in transmission rates and traffic levels within in-vehicle networks (IVNs), thereby necessitating changes in the electrical/electronic (E/E) architecture and the emergence of next-generation IVNs. This paper explores the adoption of zonal architecture with an Ethernet backbone as the vehicle topology and analyzes the factors influencing end-to-end latency. Furthermore, to evaluate the impact of IVN latency on safety-critical applications, we adopted the lane-keeping assistance system (LKAS) and employed the widely used metric, lateral error distance, to analyze how much the vehicle deviates from its intended position. We determined the feasibility of LKAS support by establishing vehicle-specific lateral distance thresholds, as allowable lateral error distances vary depending on vehicle size and comparing them with the lateral error distance. Since LKAS demands higher resolutions to achieve enhanced accuracy, this study examines the required resolution for vehicles equipped with next-generation architectures. Additionally, the paper proposes guidelines for the compression ratio of camera sensors at various resolutions and determines the maximum lateral vehicle speed achievable.