Distributed extended state observer design for strict-feedback nonlinear leader system

• Published in: ISA transactions Vol. 163; pp. 120 - 130
• Main Authors: Lv, Jixing, Wang, Changhong, Kao, Yonggui
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.08.2025
• Subjects: Distributed estimation; Distributed extended state observer; Leader-following control; Prescribed-time stability; Strict-feedback nonlinear system; Distributed estimation; Distributed extended state observer; Strict-feedback nonlinear system; Leader-following control; Prescribed-time stability
• ISSN: 0019-0578; 1879-2022; 1879-2022
• DOI: 10.1016/j.isatra.2025.05.015

Knowing the leader’s state and dynamics is crucial for leader-following control. This article addresses the distributed state/uncertainty estimation problem for a strict-feedback nonlinear leader system under directed communication topologies. The leader is characterized by Hölder-growing nonlinearities and matched uncertainty. A prescribed-time distributed estimation scheme composed of two distributed extended state observers (DESOs) is proposed. Each follower (observer node) receives only one-dimensional output estimates from its neighbors and at most one-dimensional output from the leader system, effectively reducing the communication load. First, a prescribed-time DESO (PTDESO) is proposed so that each follower can reconstruct the leader’s state and uncertainty at a time tightly prescribed by a single parameter, uniform to the initial conditions. Then, a high-gain DESO (HGDESO) is constructed, which achieves asymptotic convergence and maintains the observation errors in a small neighborhood of the origin after the prescribed time. Sufficient conditions for guaranteeing the convergence of the two DESOs are established. Ultimately, practical examples involving multiple manipulators and marine surface vehicles are provided to demonstrate the effectiveness of the proposed observers. •A class of strict-feedback nonlinear leader systems subject to matched lumped uncertainty and the Hölder growth conditions is considered.•A prescribed-time distributed estimation scheme composed of two distributed extended state observers is proposed.•Fast observation, high steady-state estimation precision and low peaking errors can be simultaneously realized.•The proposed scheme does not require the global information of the uncertainty bound and the information transmitted among followers is only one-dimensional estimated output.