Application of optimal control theory based on the evolution strategy (CMA-ES) to automatic berthing (part: 2)
Application of optimal control theory based on the evolution strategy (CMA-ES) to automatic berthing (part: 2)
复制标题
基于进化策略的最优控制理论（CMA-ES）在自动靠泊中的应用（二）
DOI：
10.1007/s00773-020-00774-x
复制
发表时间：
2020
影响因子：
2.6
通讯作者：
Naoya Umeda
中科院分区：
文献类型：
--
作者：
Maki Atsuo;Akimoto Youhei;Naoya Umeda
目前，与汽车制造相关的大多数领域都在加速自动化研究。在船舶自主作业领域，由于船舶低速机动模型的非线性、船舶低速控制能力差以及船舶与泊位发生碰撞的危险，船舶的自动靠泊一直是一个难题。以往对离线自动靠泊的研究表明，频繁切换螺旋桨旋转方向是不现实的，在柴油机情况下必须防止。为了克服这一问题，本文引入并优化了螺旋桨旋转方向切换时间的新变量。通过对目标函数的增强，建立了一种更鲁棒的离线控制方法来优化所需时间和最终船舶姿态。由于本文所提供的方法仍然是一种离线控制方法，因此其适用于在线控制必须考虑外力的不确定性和建模误差。然而，本研究得到的最优控制输入和轨迹集可作为在线预测控制建模的初始候选者。此外，该轨迹可以作为路径跟踪问题中的期望路径。
Currently, accelerated research on automation is involved in most fields associated with vehicle manufacturing. In the area of autonomous operation of marine vehicles, automatic berthing remains a problem due to the nonlinearity of the low-speed maneuvering model, poor control of the vessel at low speed, and collision with berth danger. A previous study on off-line automatic berthing reveals frequent switching of the propeller revolution direction, which is impractical and must be prevented in case of a diesel engine. In this study, to overcome this problem, a new variable for the time to switch the propeller revolution direction is introduced and optimized. Also, we establish a more robust off-line control method for optimizing the required time and the final ship attitude by enhancing the objective function. Since the method provided here is still an off-line control method, its applicability to on-line control must consider the external force uncertainty and modeling error. The set of the optimal control input and trajectory obtained in this study, however, are applicable as an initial candidate for on-line predictive control modeling. Furthermore, the trajectory can serve as the desired path in a path-tracking problem.