Closed-Loop Identification and Control of Inland Vessels

Abstract

Extensive research has been conducted on a navigation system for inland vessels at the University of Stuttgart and at the Max Planck Institute for Dynamics of Complex Technical Systems, Magdeburg (Focus Max Planck Gesellschaft, Computer at the helm, http://www.mpg.de/942027). As part of this navigation system, a model-based track-keeping controller has been developed. A high performance controller is required because of the reduced space available on rivers and canals. The control structure consists of two components, a feedback and a feedforward block, where the former is provided by a linear quadratic gaussian controller. Both components require the ship dynamics model and thus, the parameter estimation of the underlying model is a key issue to achieve high performance. In this chapter, we firstly consider Monte Carlo simulations to generate data of the closed-loop system and then the parameters of a continuous-time steering dynamics model are identified. The parameter estimation problem is solved applying an instrumental variable method, which takes into account the control structure. Parameter identification using real closed-loop experiments is also considered. Additionally, we evaluate the experiments for parameter estimation through a sensitivity analysis.

Transport of Water versus Transport over Water

“…• Modeling (for control): Closed-loop identification, data assimilation, Kalman filters, probabilistic maps, potential fields, dynamical models, hydraulic models, LTV models [2,9,13,15] • Different control techniques:…”