Multi-rate Control System Design for Distributed Parameter Systems via Induced ℒ 2 Mixed Sensitivity Optimization

Richard P. Metzger, Delano R. Carter, Armando Rodriguez

Research output: Chapter in Book/Report/Conference proceedingConference contribution


The problem of designing near-optimal finite-dimensional controllers for stable multiple-input multiple-output (MIMO) distributed parameter plants under multi-rate sampled-data control is discussed in this paper. A weighted ℋ -style mixed-sensitivity measure penalizing the control is used to define the concept of optimality. Controllers are found by solving a finite-dimensional sampled-data optimization. A priori computable conditions for the approximants are provided such that the resulting finite-dimensional controllers stabilize the multi-rate sampled-data controlled distributed parameter plant and are near-optimal. The proof is dependent upon a key fact that the control input is appropriately penalized in the optimization. Another assumption of the presented technique is the plant may be approximated uniformly by finite-dimensional systems. It is shown how the optimal performance may be approximated to any desirable level of accuracy by solving a single finite-dimensional problem using a suitable finite-dimensional approximant. No infinite-dimensional spectral factorizations have been utilized. This paper provides a straight forward approach of control design for a large class of MIMO distributed parameter systems under multi-rate sampled-data control.

Original languageEnglish (US)
Title of host publicationProceedings of the American Control Conference
Number of pages6
StatePublished - 2003
Event2003 American Control Conference - Denver, CO, United States
Duration: Jun 4 2003Jun 6 2003


Other2003 American Control Conference
Country/TerritoryUnited States
CityDenver, CO

ASJC Scopus subject areas

  • Control and Systems Engineering


Dive into the research topics of 'Multi-rate Control System Design for Distributed Parameter Systems via Induced ℒ 2 Mixed Sensitivity Optimization'. Together they form a unique fingerprint.

Cite this