Mixed Traffic Scheduling of URLLC Industrial Automation Under Network Calculus Analysis

Ultra-reliable and low-latency communications (URLLC) holds significant potential in the wireless industrial automation, especially for increasing flexibility while guaranteeing stringent quality-of-service (QoS) requirements. Inspired by the traffic shaper in time-sensitive networking (TSN), this paper proposed a mixed traffic scheduling scheme in URLLC-enabled industrial automation, where time-triggered (TT) traffic and event-triggered (ET) traffic are separated into different queues and shaped by the gate-control lists (GCLs) and the credit-based shaper (CBS). Due to the preemption of TT traffic, the end-to-end (E2E) latency of ET traffic may exceed its requirement. To characterize the effect of TT traffic on the latency of ET traffic, we derive the delay violation probability bound for ET traffic using network calculus with the guarantee of transmission of TT traffic. On the other hand, since URLLC requires conservative resources to ensure high reliability within short latency, it is challenging to achieve bandwidth-saving scheduling for mixed traffic. To minimize the total downlink bandwidth while ensuring the QoS requirements, a two-step method is proposed to find the global optimal solution of the send slope and idle slope of CBS, bandwidth allocation, subchannel assignment and preemption duration. Simulation results validate our analysis and show the bandwidth-saving performance gain by optimizing CBS parameters and PHY layer resources with the proposed TSN-based scheduling scheme.