导航

ACTA AERONAUTICAET ASTRONAUTICA SINICA ›› 2022, Vol. 43 ›› Issue (3): 325209-325209.doi: 10.7527/S1000-6893.2021.25209

Previous Articles     Next Articles

A scheduling table generation method for time-triggered flows based on importance sampling

LUO Zexiong, QU Guoyuan, YAN Long, TANG Xueqian   

  1. Department of Electronics, Chinese Aeronautical Radio Electronics Research Institute, Shanghai 200233, China
  • Received:2021-01-05 Revised:2021-02-24 Online:2022-03-15 Published:2021-04-29

Abstract: Time-triggered Ethernet (TTE) is the backbone interconnections in the distributed integrated modular avionics systems, with a congestion-free time-triggered (TT) scheduling table being needed to be generated.Although the formal method using SMT can describe multi-typed scheduling constraints abstractly, more complicated constraints could reduce the efficiency of scheduling generation due to the existence of job chain dependencies among TT flows and between TT flow and partition scheduling.Importance sampling (IS) is used for a kind of heuristic TT scheduling generation, in which the possibility density distribution (PDF) of start trying offsets is twisted by experimental PDFs according to statistical values of the start-trying offsets using the "As Soon As Possible" (ASAP) method.IS-based scheduling table generation is divided into two stages to obtain more feasible scheduling and more timely job chains, respectively.The case study shows that in the first stage, a set of distribution for start-trying offsets is gained to make TT flows easy to be strictly periodically scheduled; in the second stage, optimal offsets are conducted to make shorter job chains delays according to the "best of the worst" criterion, and IS is applicable for the scenarios with or without synchronization between time-triggered network and partition-scheduled processing modules.

Key words: avionics, time-triggered scheduling table, time-triggered Ethernet, importance sampling, job dependency

CLC Number: