Report Title: Improved Parameter Estimation of Dynamic Systems with Kolmogorov–Zurbenko Filtering
Reporter: Professor Wang Qingguo
Report Time: 14:00, December 2
Report Venue: Room A507, School of Mechanical and Electrical Engineering
Hosted by: School of Mechanical and Electrical Engineering
Reporter’s Profile:
Professor Wang Qingguo is a Fellow of the Academy of Science of South Africa. He currently serves as a Chair Professor at BNU-HKBU United International College and a Professor at the Institute of Artificial Intelligence and Future Networks, Beijing Normal University. In 1987, he earned his Ph.D. in Industrial Automation from Zhejiang University and received the title of "Outstanding Graduate." He held positions as a Full Professor in the Department of Electrical and Computer Engineering at the National University of Singapore and a Distinguished Professor at the Institute of Intelligent Systems, University of Johannesburg, South Africa. His academic focus lies in automation and artificial intelligence, with primary research interests in modeling, estimation, prediction, optimization, and control of complex systems. He has published over 500 papers in international journals and authored seven academic monographs published by Springer. His work has been cited nearly 23,000 times, with an H-index of 82. In 1990, he received the "Youth Science and Technology Award" from the China Association for Science and Technology and was recognized as an "Outstanding Ph.D. Graduate" by the State Education Commission. From 1990 to 1992, he was awarded the Alexander von Humboldt Research Fellowship in Germany. In 2011, he was granted the "Most Cited Paper Award" for the years 2006–2010 by the prestigious journal Automatica. He was listed as a Highly Cited Researcher by Thomson Reuters in 2013 and received the "Most Influential Paper Award" in 2014 for the 30th anniversary of the journal Control Theory and Applications. In 2020, he was ranked among the top 2% of scientists worldwide for "Career-Long Scientific Impact" and "Annual Scientific Impact" in a list released by Stanford University. He has served as the Chair of the IEEE Singapore Control Chapter four times and as the General Chair of the Asian Control Conference and several IEEE international conferences. Currently, he is the Deputy Editor-in-Chief of the internationally renowned journal ISA Transactions. He has supervised approximately 40 doctoral students and 30 postdoctoral fellows.
Report Overview： System identification involves constructing models of dynamic systems from measured data. A key challenge is achieving accurate estimation of model parameters despite noisy data, which often arises from sensor inaccuracies and communication errors. The least squares estimation leads to inconsistent parameter estimates when the system is disturbed by correlated noise. To mitigate the impact of measurement errors on parameter estimation, this paper uses the Kolmogorov–Zurbenko (KZ) filter and proposes a consistent parameter estimation method. The KZ filter effectively preserves long-term trends while suppressing high-frequency noise through its sharp frequency cut-off. In the proposed method, the input and output signals are first filtered simultaneously using the KZ filter, and then the filtered signals are used for parameter estimation. Three linear systems are employed to validate the efficiency of the method. The simulation results show that the proposed method achieves the lowest parameter estimation error and prediction error in all cases compared to the benchmark methods.
