Neural network based adaptive finite-time distributed estimation for an uncertain leader

This paper investigates a two-step finite-time distributed estimation scheme for an uncertain leader. Unlike the previous achievements, the leader with unknown nonlinearity and uncertain input is considered, and the whole scheme is fully distributed and output-based. Firstly, a local neural network (NN) finite-time observer is proposed to estimate the unavailable states / uncertain dynamics of the leader, where the NN is used to approximate the uncertain dynamics. Then, based on the local interaction among agents, an NN finite-time distributed observer is devised for all the followers to reconstruct the system states / NN weights broadcasted by the local observer. By utilizing a combination of the local and the distributed observer, the unavailable states and the uncertain dynamics of the leader can be reconstructed by each follower in a finite time. Finally, simulation examples are presented to demonstrate the validity of our scheme.