An Accuracy-enhanced Delay Upper Bound Model for In-vehicle Time-Sensitive Networking with Network Calculus

Time-sensitive networking (TSN) is emerged as a prominent solution for in-vehicle network (IVN) because of the deterministic communication capability in multi-traffics transmission. However, it is complicated to configure the gate control lists (GCLs) among multiple TSN switches under massive heterogeneous traffics, especially in IVN where communication and computing resources are seriously limited. Therefore, a new algebraic deterministic network calculus (DNC) algorithm named as fine-grained left-over service curve for time-aware shaper (FLSC-TAS) is proposed to evaluate more accurate delay upper bounds for time-critical traffics, aiming to simplify the GCL configurations for in-vehicle TSN. The proposed FLSC-TAS enhances the accuracy of service curve and traffic arrival curve considering the characteristics of priority queues in TAS mechanism. Compared with two commonly used baseline methods, the results demonstrate that the proposed FLSC-TAS algorithm achieves higher accuracy in delay upper bound evaluation. Furthermore, the average runtime is reduced by 14.04% compared to the most commonly used method.