Robust State Estimation and Control for Uncertain Systems withRandom Structure.

---

Investigator(s): Dr Valeri Ougrinovski (valu@eeadfa.edu.au) Phone: 02-62688219

Complex systems consisting of continuous-time plant controlled by a discrete-time events have been extensively studied since 1960's; e.g.. This class of automatic control systems has been called upon to address issues of reliability for engineering controlsystems in the face of randomly occurring hardware failures and other
abrupt changes in the system structure. Abrupt changes in the system structure are typical for many industrial systems such as electric power systems, communication systems, manufacturing systems, aircraft flight control systems,military applications such as multitarget tracking, computer networks. For example, randomly occurring information traffic jams experienced by onecomputer in a distributed computer network can significantly compromise performance of the network. In control engineering, systems with random structure are used in designing fault tolerant control systems which are able to automatically detect failures and then
reconfigure the controller in response to the changes in the system structure.

One possible way of modelling such systems is to model the abrupt changes in the system structure usingdiscrete-state Markov processes. This approach leads to what is called Markov Jump Parameter system (MJP). The aim of this research is to investigate a wide range of robust control MJP systems. We focus on research into new possibilities that such investigation may offer for robust state estimation and control of systems with random structure. The theory resulting from this research makes it possible to obtain tractable solutions to a number of important robust state estimation and control problems involving uncertain systems with MJP.

 

Research Activities of the School

 

---