.png){kind=icon}
.png){kind=icon}
Return
Data-driven model predictive control for underactuated USV path tracking with unknown dynamics
.png){kind=icon}
PRE
AI
DOI：10.1016/j.oceaneng.2025.121457.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper presents a novel model predictive control (MPC) path tracking method based on dynamic fuzzy neural network (DFNN) modeling for unmanned surface vehicles(USVs) with unknown dynamic parameters. Although MPC is widely used in USV control, previous studies have typically relied on the known dynamic models of USVs. To the best of our knowledge, this is the first time such an approach has been applied to real-time modeling and control of USVs. In practical scenarios, many parameters in the USV model are difficult to obtain, representing one of the inherent challenges of the MPC algorithm. In this paper, we develop a fundamental dynamical model for the underactuated USV using a DFNN, and the MPC algorithm constructs a cost function based on this model to solve for the optimal control variables. This approach enables the MPC to determine control variables without relying on explicit parameters of the mathematical model. Additionally, the DFNN undergoes online training in a data-driven manner, enhancing its accuracy as a model. To validate the effectiveness of the proposed algorithm, we conducted a series of simulation experiments focused on the path tracking control problem of an underactuated USV. These results were then compared and validated against current mainstream control algorithms. The numerical simulation results demonstrate that the MPC-DFNN algorithm offers satisfactory control performance, even in the presence of unknown USV dynamic parameters.
Keywords:
Underactuated USV
Model predictive control
Dynamic fuzzy neural network
Modeling
Path tracking
Journal
IF:
5.5
Papers:
5.7K
Citations:
7.6W
Organization
No organization information available