Multivariate Modeling and Adaptive Control of Autonomous Ferries

Abstract

This paper deals with the design and evaluation of a multivariate-based adaptive controller for marine surface vessels. The goal is to make a system that periodically updates the vessel model, used by a dynamic positioning (DP) controller, in order to improve the motion-control performance. The model is generated by analyzing velocity and acceleration data using multivariate analysis (MVA) to fit a 3 degrees of freedom (DOF) surface vessel model. Full-scale experimental results with the prototype autonomous passenger ferry milliAmpere show that the adaptive control system works as intended, and manages to make significant improvements in all performance metrics compared to a non-adaptive control system. The results also show that the system is robust even with the impact of unmeasured wind and wind gusts.