Autonomous docking using direct optimal control

Abstract

We propose a method for performing autonomous docking of marine vessels using numerical optimal control. The task is framed as a dynamic positioning problem, with the addition of spatial constraints that ensure collision avoidance. The proposed method is an allencompassing procedure for performing both docking, maneuvering, dynamic positioning and control allocation. In addition, we show that the method can be implemented as a real-time MPC-based algorithm on simulation results of a supply vessel.

The review unmanned surface vehicle path planning: Based on multi-modality constraint

“…Some scholars have done some research on this topic, such as Djouani & Hamam (Djouani, K., & Hamam, Y. 1995), Okazaki, T., et al (Okazaki, T., et al, 2000), Meyer, P. J., et al (Meyer, P. J., et al, 2019) and Martinsen, A. B., et al (Martinsen, A.…”

Optimization-Based Automatic Docking and Berthing of ASVs Using Exteroceptive Sensors: Theory and Experiments

“…Automatic docking is a complex problem, which includes planning and performing maneuvers to control a vessel to a desired orientation and position, while adhering to spatial constraints in order to avoid collisions. In order to perform the docking, we expand upon [32], [33], where we use a docking trajectory planner, constructed as an Optimal Control Problem (OCP). This allows the planner to take into account the vessel dynamics in terms of a mathematical model, as well as the harbor layout in terms of a map of landmasses.…”

“…The docking method from [32] is a nonlinear model predictive controller (NMPC) that takes into account vessel dynamics in the form of its dynamic model, as well as collision avoidance by planning trajectories within a convex set, based on the harbor layout. Advantages of this approach include the explicit handling of static obstacles, the planning of dynamically feasible trajectories, and a flexible behavior shaping via the nonlinear cost function.…”

“…Additionally, due to the non-convexity of the optimal control problem, guarantees on run time or feasibility are not provided. In this paper, we build on [32], [33] and propose a novel algorithm for dynamically creating a convex safety set, based on a map of the environment. We also show how this method can be combined with sensor data from on board sensors such as LIDAR pointclouds, and ultrasonic distance sensors, in order to account for missing or inaccurate map data.…”

“…The contributions of this work can be split into two main categories, namely methodology, and implementation. 1) Methodology builds on the work in [32], with the following improvements:…”