Event-triggered extended state observers design for dynamic positioning vessels subject to unknown sea loads

Abstract

This paper presents a resource-aware state and uncertainty estimation method for dynamic positioning vessels subject to uncertain kinetics and unknown sea loads induced by wind, waves and currents. Specifically, an event-triggered extended state observer (ESO) is proposed such that unnecessary communications are avoided and sensor-to-observer communication cost are drastically reduced. A realizable event-triggering condition is established depending on the position-heading measurements only. It is proven that the observation errors are input-to-state stable (ISS), and Zeno behavior will not occur. Aimed at achieving finite-time convergence, an event-triggered finite-time convergent ESO is further designed. By resorting to a homogeneous Lyapunov function, the estimation errors are proven to be finite-time ISS. A salient feature of the proposed event-triggered ESOs is that a joint observation of the unmeasured velocities, unknown kinetics and unknown sea loads can be achieved with reduced communication cost. The proposed event-triggered ESOs can be applied to various motion scenario such trajectory tracking, path following, target tracking, and even formation control of marine vessels. Simulation results show that the communication times are largely decreased compared with the time-triggered ESOs.