Design a Drift Path for Unmanned Surface Vehicles using Nonlinear Model Predictive Control

With the advancement of unmanned driving technology, Unmanned Surface Vehicles (USVs) have become a significant topic in marine exploration, environmental monitoring, and military applications. USVs are characterized by their efficiency, flexibility, and remote-control capabilities, playing an increasingly important role in surface operations. As the number of USVs increases, the issues of automated charging, maintenance, and management will become substantial time and labor resource burdens. However, the automatic docking of USVs faces numerous technical challenges. Due to the differences in driving mechanisms, inertia, and resistance between vessels and vehicles, the kinematic principles applicable to vehicles are not suitable for small boats, making it impossible to employ parallel parking methods used for vehicles directly. Consequently, using drift docking technology for USVs has emerged as a critically important solution, enabling fully automated management of unmanned fleets, including automatic cruising, docking, and charging, thereby reducing labor and time-related maintenance costs. This paper uses nonlinear model predictive control and the Gazebo platform as the algorithm and simulation environment for USV drift parking. A dual-propeller USV controller is designed using a simplified kinematic-dynamic bicycle model. The simulation results show that this method can successfully operate on the Gazebo platform and has the possibility of being implemented in an environment where unmanned ships sail in real waters.