Neural Modified Extended State Observer for Autonomous Maritime Transportation Vehicles with Actuator Faults and Unmatched Nonlinearities

Abstract

This paper investigates extended state observer design problem for autonomous maritime transportation vehicles consisting of multiple unmanned aerial vehicles (UAVs) and unmanned surface vehicles (USVs) via the cooperation–competition network. Through two coordinate transformations, the dynamic model of UAVs and USVs is transformed into a standard second-order heterogeneous multi-agent system in unified earth-fixed frame. In order to recover the unavailable velocities, unpredictable fault, total disturbance as well as to estimate the heterogeneous match/unmatched nonlinearities, neural network (NN) is incorporated into the design process of the modified distributed extended state observer (MESO). Based on the Lyapunov stability analysis, it is proved that all the error signals are uniformly ultimately bounded (UUB) and the bounds could be arbitrarily small by choosing appropriate parameters. Simulation results verify the effectiveness of the proposed method.