Robust Nonlinear Model Predictive Control of an Autonomous Launch and Recovery System

Launching and recovering a lifeboat from a mother ship is a critical task for rescuing people in high sea states, which can be dangerous to both the mother ship crew and lifeboat personnel. A reliable and efficient control system is crucial to reducing the risk but has not been developed to a mature stage to establish an autonomous launch and recovery system (LARS). A successful manually controlled launch and recovery (L&R) mission relies on empirically assessing the risk and planning the operation ahead of initiating the process. This article proposes a control scheme for the LARS which executes the task in two stages: the L&R risk assessment is conducted in the first stage before hoisting the lifeboat; then in the second stage, input signals are manipulated to accomplish the task once the mission is identified to be safe. We propose a robust tube-based model predictive control (TMPC) law in both stages. It can explicitly consider uncertainties in the LARS model and guarantee constraint satisfaction by bounding possible system trajectories in a predefined tube. Hence degradation of control performance caused by inaccurate system modeling can be minimized to improve the operation safety level of the entire process. The performance of the proposed control scheme is demonstrated by numerical simulations.