导航

Acta Aeronautica et Astronautica Sinica ›› 2023, Vol. 44 ›› Issue (18): 328125-328125.doi: 10.7527/S1000-6893.2023.28125

• Electronics and Electrical Engineering and Control • Previous Articles    

Online time-triggered scheduling method based on publish/subscribe architecture

Yifan GAO, Feng HE(), Sifan YU   

  1. School of Electronics and Information Engineering,Beihang University,Beijing  100191,China
  • Received:2022-10-18 Revised:2022-11-23 Accepted:2023-03-07 Online:2023-04-03 Published:2023-03-31
  • Contact: Feng HE E-mail:robinleo@buaa.edu.cn
  • Supported by:
    National Natural Science Foundation of China(62071023)

Abstract:

The scheduling solution of hybrid critical messages is the key to its application in avionics systems, and increasingly dynamic applications rely more on the online solution of time-triggered scheduling. At present, time-triggered network scheduling is mostly based on offline scheduling design, which take a long time to be generated and are difficult to be adjusted online after generation in large-scale networking applications. In order to solve the scheduling table faster and make online adjustment, this paper constructs a time-triggered network architecture model based on the publish/subscribe model. Then, an online time-triggered scheduling algorithm is proposed based on unified time slots. Time is discretized into time slots with the unified maximum length, which optimizes the scheduling solution space and reduces the time of scheduling table generation. The link load is measured by the time slot length and is balanced in the process of scheduling to reduce the end-to-end delay of rate-constrained messages. The experimental results show that for the network with 300 messages, the scheduling solution speed of the proposed algorithm is thousands of times that of the SMT algorithm, and the worst-case end-to-end delay of rate-constrained messages is reduced by 17.4% compared with that of the SMT algorithm. For the network with 2 000 time-triggered messages, the time of generating schedule table with the proposed method is on the order of 100 milliseconds.

Key words: time-triggered, publish/subscribe architecture, online scheduling, time slicing, real-time performance

CLC Number: