Engineering HDR Seminar Series 11, 2012: Development of Common Observer Schemes using Sliding Mode Observers for Robust State Estimation and Fault Detection and Isolation

Mr Ng Jiunn Yea, RAM postgraduate student

Date: 2012-07-12
Time: 15:00 to 16:00
Venue: Engineering meeting room 1, 5-4-22


In controlling dynamic engineering systems for optimal performance, the estimates of states are very useful as direct measurements of those states are impractical or at times may be impossible. The reasons for this impracticability are due to high cost of sensors and sensors may also suffer from inherent error such as noise. In addition, the physical surroundings may not even permit the installation of a sensor. An observer provides a solution as it uses the measured outputs and inputs of a system, together with a dynamic mathematical model of the system to estimate the states. However, the mathematical model is usually a representation of the system at only a certain operating point; at a different operating point the mathematical model changes and could cause the observer to give an inaccurate state estimation, or even worse raise a false alarm during a fault-free situation, or hide the effect of the fault during the faulty situation. An interesting solution to these problems is the common observer, which essentially is a bank of observers that can provide an accurate state estimation for various mathematical models. Preliminary research includes the analysis of a common observer scheme in the literature, as well as the investigation of its existence conditions.

About the Speaker

Mr Ng Jiunn Yea graduated with Bachelor of Engineering (Mechatronics) with Honours from Monash University in 2009. He joined Monash University for postgraduate studies in 2011 under the supervision of Dr Edwin Tan Chee Pin. Prior to his postgraduate studies, he worked as an Equipment Engineer in Flextronics for 1.5 years. His research in Monash University focused on control systems. The present study aims to develop common observer schemes for robust state estimation and fault detection and isolation.