Analogies between Markov jump and switched systems: applications to networked control

• Grant number: 11/12871-0
• Support Opportunities: Regular Research Grants
• Duration: December 01, 2011 - November 30, 2014
• Field of knowledge: Engineering - Electrical Engineering - Industrial Electronics, Electronic Systems and Controls
• Principal Investigator: Alim Pedro de Castro Gonçalves
• Grantee: Alim Pedro de Castro Gonçalves
• Host Institution: Faculdade de Engenharia Elétrica e de Computação (FEEC). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil

Abstract

We want to study Markov Jump Linear Systems and their relationship to switched systems where the commutation function is the control input. For markov jump systems, if the Markov chain state, often called mode, is available to the controller it is called mode-dependent. If the Markov variable is not available we have mode-independentcontrol. Recent results for linear parameter varying systems suggest we can obtain amode-independent state feedback control based on a switching rule. Such a control mightprobably by less conservative than the available designs from the current literature. Wewill also use Markov jump systems to model networked control systems. We are particularly interested in two aspects, namely, the models for the occurrence of transmissionfailures and the strategy to be taken whenever there is a packet loss. (AU)