INTELLIGENT LEARNING CONTROL OF MARINE VESSELS UNDER HARSH ENVIRONMENTS

Abstract

The objective of this research is to develop reliable DP tracking system for deepsea accommodation vessels to equip them with the capability to operate in harsh environmental conditions. Firstly, in order to reduce wear and tear on the thrusters due to the inclusion of high frequency wave-induced motion, a neural network (NN) based adaptive filter is introduced to separate the low-frequency and high-frequency motion in the presence of time-varying sea condition and model uncertainty. Secondly, given a set of filtered low-frequency motion, the accommodation vessels should be controlled to serve other marine platforms and structures by synchronously moving along them under different marine environments. To guarantee smooth and safe operation, barrier Lyapunov methods, supervisory control as well as artificial potential based robust control are investigated to achieve the control objective from different perspectives. Finally, in order to ensure the control performance and lower energy consumption simultaneously, an optimized backstepping approach is proposed and utilized to the marine vessel through reinforcement learning.