Approach to Improving Asynchronous Traffic Shaping Performance Using a Combination of Shaper

In recent vehicles, the complexity of Electrical/Electronic ($\mathbf{E} / \mathrm{E}$) architectures and traffic usage has increased due to adopting technologies such as Advanced Driver-Assistance Systems (ADAS) in vehicles. Accordingly, The IEEE 802.1 Time-Sensitive Networking (TSN) Task Group (TG) released the TSN to ensure Ethernet’s time determinism and time synchronization-based Time-Aware Shaper (TAS) algorithms have been actively studied. However, TAS algorithms based on time synchronization suffer from degraded real-time performance when time synchronization fails and increased traffic load for time synchronization. The IEEE 802.1 TSN TG is considering an IEEE 802.1 Qcr is called Asynchronous Traffic Shaping (ATS), and it is based on Token Bucket Emulation (TBE) derived from Urgency-Based Scheduler (UBS). This standard has yet to be established, and various approaches are being studied. According to existing studies, Length Ratio Quotient (LRQ), TBE, and ATS algorithms were used as asynchronous traffic control methods, and only one algorithm was applied to a single system. In this paper, we propose three approaches to combining shapers based on the performance analysis of each LRQ, TBE, and ATS algorithm. We verify their performance by measuring the proposed methods’ end-to-end delays of each traffic. The results show that the approach that combined the ATS and LRQ shaper and applied the LRQ algorithm to all incoming frames with a size smaller than a threshold (42 bytes) has the most stable performance among the proposed approaches. In addition, iterative experiments confirm this approach’s reproducibility.