In this talk, we introduce two novel observer design tools for nonlinear control systems, i.e., contractingobserver and Parameter Estimation-Based Observer (PEBO). First, we propose an approach to designglobally convergent reduced-order nonlinear observers via contraction analysis and convex optimisation.Despite the fact that contraction is a concept naturally suitable for state estimation, the existing solutionsare either local or relatively conservative when applying to physical systems. To address this, we showthat this problem can be translated into an offline search for coordinate transformation after which thedynamics is transversely contracting. In the second part of the talk, we present a new observer tool calledPEBO, whose main idea is translating the state estimation problem into one of estimation of constant,unknown parameters. We present also PEBO in a unified framework in a unified framework together withby now classical–Kazantzis-Kravaris-Luenberger (KKL) and Immersion and Invariance (I&I) observers.At the end of the talk, we extend the PEBO from Euclidean space to matrix Lie groups, which is thenapplicable to the widely popular problem of visual inertial SLAM in the robotics community.
Bowen Yi obtained his Ph.D. degree in control engineering from Shanghai Jiao Tong University, Chinain 2019. From 2017 to 2019 he was a visiting student at Laboratoire des Signaux et Systèmes, CNRS-CentraleSupélec, Gif-sur-Yvette, France. From September 2019 to September 2022, he was PostdoctoralResearch Associate in Australian Centre for Field Robotics, The University of Sydney, NSW, Australia.Currently, he is Research Fellow at Robotics Institute, University of Technology Sydney, NSW, Australia.His research interests involve the algorithms for estimation, learning and control of nonlinear systems,with special emphasis to robotics. He received the CCTA Best Student Paper Award from the IEEEControl Systems Society in 2019 for his work on nonlinear observer design.
B. Yi, C. Jin and I. R. Manchester, Globally convergent visual-feature range estimation with biased inertial measurements. Automatica , 146, 110639.
B. Yi, C. Jin, L. Wang, G. Shi and I. R. Manchester, "An almost globally convergent observer for visual SLAM without persistent excitation," 2021 60th IEEE Conference on Decision and Control (CDC), 2021, pp. 5441-5446, doi: 10.1109/CDC45484.2021.9683624.
B. Yi, R. Ortega and W. Zhang, "On State Observers for Nonlinear Systems: A New Design and a Unifying Framework," in IEEE Transactions on Automatic Control, vol. 64, no. 3, pp. 1193-1200, March 2019, doi: 10.1109/TAC.2018.2839526.
B. Yi, R. Wang and I. R. Manchester, "Reduced-Order Nonlinear Observers Via Contraction Analysis and Convex Optimization," in IEEE Transactions on Automatic Control, vol. 67, no. 8, pp. 4045-4060, Aug. 2022, doi: 10.1109/TAC.2021.3115887.