In-Vehicle Network Average Response Time Analysis for CAN-FD and Automotive Ethernet

Abstract

In-vehicle network (IVN) technologies are changing corresponding to the rapidly increasing demand in autonomous driving and vehicle electrification. Due to the increase in real-time data generated from the control modules and sensors, controller area networks (CANs) have a heavier burden to satisfy the strict delay bounds. Therefore, CAN with flexible data rate (CAN-FD) and Automotive Ethernet (AE) have drawn attention as promising IVNs. This paper provides an overview of the IVN transmission protocols, time delay characteristics, and analysis methods. Most of the existing IVN timing analysis studies assume the worst case and consider only the occurrence of periodic or aperiodic messages, however for more accurate control, the average time delay needs to be accurately estimated. In this paper, CAN, CAN-FD, and AE analytical models based on queueing analysis are proposed to estimate the average response time for mixed traffic. For CAN and CAN-FD systems, the network synchronization statuses are considered and AE with 1-level preemption to support deterministic delays is analyzed. The improved accuracy of the newly derived analytical models is verified based on comparison with existing CAN queueing models through simulation experiments.