Maximally Permissive Distributed Control of Large Scale Automated Manufacturing Systems Modeled with Petri Nets

Hesuan Hu, Yang Liu, Mengchu Zhou

Research output: Contribution to journalArticlepeer-review

38 Scopus citations

Abstract

Ensuring nonblockingness remains challenging for automated manufacturing systems (AMSs) owing to their discrete event dynamics. Both scalability and maximal permissiveness are essential for the synthesis and implementation of their centralized supervisors. Inspired by the divide and conquer philosophy, this brief proposes a partition methodology and distributed control technique for large-scale AMSs. They are represented as interconnected and overlapping subsystems sharing some common components in terms of buffers. For each subsystem, a local supervisor is designed based on its local behavior and neighboring information only. Generalizing the existing results, we develop a condition under which the control law via decomposition promises the maximal permissiveness. Buffer capacities are well designed for the sake of their decomposition into multiple overlapping subsystems. Theoretical results are developed to characterize the behavior compatibility among local controllers. An experimental study illustrates the effectiveness of the proposed method.

Original languageEnglish (US)
Article number7029080
Pages (from-to)2026-2034
Number of pages9
JournalIEEE Transactions on Control Systems Technology
Volume23
Issue number5
DOIs
StatePublished - Sep 1 2015

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • Electrical and Electronic Engineering

Keywords

  • Automated manufacturing systems (AMSs)
  • Petri nets (PNs)
  • deadlock prevention
  • discrete event systems
  • distributed control
  • maximal permissiveness
  • supervisor synthesis

Fingerprint

Dive into the research topics of 'Maximally Permissive Distributed Control of Large Scale Automated Manufacturing Systems Modeled with Petri Nets'. Together they form a unique fingerprint.

Cite this