Continuous-time system identification of the steering dynamics of a ship on a river

Abstract

In this study, we consider the parameter estimation problem of a ship dynamics model. We consider two possible approaches to identify a continuous-time model from real data obtained on a river, where the presence of disturbances is a key issue. The first approach is identification through optimisation using a disturbance observer. The second approach corresponds to the refined instrumental variable method for linear parameter varying systems. In addition, we evaluate the accuracy of the parameter estimation through a sensitivity analysis. The obtained results show an improvement in the parameter estimates compared to identification procedures that do not consider the river disturbances. The application of the model for track-keeping control is also illustrated.

Closed-Loop Identification and Control of Inland Vessels

“…The former is defined through the so called Nomoto model [28] and the latter is derived in [7] (see more in [4,29]). The steering dynamics is then defined by Hereafter we consider that for the drift model, the contribution of the rudder force is negligible, as in [7], i.e., The superscript N is used to denote normalized parameters (i.e., parameters that do not depend on the velocities v r and v abs ).…”

“…Using open-loop experiments, continuoustime (CT) identification methods have been successfully tested in [29] to cope with this issue. Nevertheless, some drawbacks of open-loop identification of inland vessels have been reported in [35], namely, the identification experiments can not be automated, and the rudder excitation might depend strongly on the skipper.…”

“…Experiments like zig-zag maneuvers [29], spiral maneuvers, and multi-frequency binary tests [8] can be used to identify the parameters of ship dynamics models. This kind of experiments are performed in open-loop, in contrast to closed-loop experiments that are done using a controller.…”

“…Process noise, measurement noise, and/or structural model uncertainties might explain such phenomenon. In this study, this parameter accuracy problem is evaluated through a sensitivity analysis, as in [29].…”

Nonlinear Modeling for a Water-Jet Propulsion USV: An Experimental Study

“…However, to estimate and identify the hydrodynamic derivatives, the forces and torque exerted on the USV need to be accurately measured. Unfortunately, this requires some strict experimental conditions that are often impossible [8].…”