Extended state observer-based non-singular practical fixed-time adaptive consensus control of nonlinear multi-agent systems

This article considers the practical fixed-time control algorithm of multi-agent systems (MASs) with unmeasurable states and disturbances. Neural networks are introduced to fit the nonlinear functions of MASs due to the existence of unknown dynamics. The unmeasured states and unknown disturbances in MASs are estimated by applying the extended state observer (ESO). The adding power integration method is combined with the backstepping approach, which avoids the singular phenomenon in the fixed time controller. Besides, the command filter is used in the standard backstepping approach, reducing the computational complexity in the controller design. On the basis of the practical fixed-time stability theory, it is guaranteed that the tracking error of MASs converges within a fixed time, and the initial states of MASs do not affect the convergence time. The simulation case verifies the validity of the proposed MASs consensus algorithm.