Research paper
Design and implementation of a testing platform for ship control: A case study on the optimal switching controller for ship motion

https://doi.org/10.1016/j.advengsoft.2023.103427Get rights and content

Highlights

  • The structure and design process of ship testing platform is described..

  • Research process of optimal switching strategy for control are summarized.

  • A combined optimal switching controller of ship motion is designed.

  • A new method is proposed to reduce the large turning error of ships during navigation.

Abstract

Fully autonomous navigation is the highest level of ship autonomy, which has high requirements for the adaptability of the ship control system. Taking navigation with large turning errors as an example, this paper combines the intelligent optimization controller with the traditional controller and adopts the controller switching strategy to guide the ship’s path following. In order to illustrate the feasibility of optimizing the switching mode of the motion controller to reduce the large turning error during ship navigation, this paper designs a test platform for the intelligent navigation of the model ship, and tests the effect of the optimal switching strategy of the controller. Firstly, the structure of the intelligent ship system is analyzed, and the framework of the intelligent navigation system is emphasized. Secondly, based on the analysis of the characteristics of controllers, the optimal switching strategy of the motion controller is proposed. Then, the construction process of the testing platform is described. Finally, the optimal switching strategy of the motion controller is verified on this test platform for the case of a large turning error of the ship. The results show that the implementation of the switching strategy can effectively reduce the control error compared with the mode without using the strategy. The experiment extends the ship motion control mode strategically and improves the adaptability of ship motion control. Based on the testing platform, this experiment can simulate the process of the control system autonomously selecting an appropriate controller according to the ship’s motion state and navigation scenarios, which is helpful to enhance the adaptability of the ship in the face of special navigation scenarios.

Introduction

The Maritime Safety Committee of the International Maritime Organization (IMO) has put forward that ships have four levels of autonomy. At present, ships have developed to the second and third levels [1]. The research content of this paper is aimed at the third level (Remotely controlled ship without seafarers on board) or the fourth level ships (Fully autonomous ship). With the integration and application of high and new technologies such as artificial intelligence, automatic control, data science and the shipping industry, to realize the full autonomous navigation of intelligent ships, many experts and scholars have studied it and designed many controllers to realize ship motion control. No matter how complex and advanced the controller is, it needs corresponding experiments to verify it. The implementation of a ship navigation experiment needs many conditions, such as building a scaled model ship, carrying intelligent equipment and building an experimental environment, which is not easy to achieve. Therefore, many studies take virtual simulation as the primary verification method. In the part of the research with experimental verification, a few articles elaborate on the structure of the whole experimental system, and the experimental methods are also different. These studies can be summarized as follows: (1) Introducing experimental data into the simulation as a reference; (2) The research object and experimental results are given directly, and the testing platform is not described in detail; (3) The intelligent ship and the construction of the testing platform are introduced, in which only hardware control system or software system is described in detail; (4) The intelligent ship and the construction of the testing platform are introduced in detail, including software and hardware systems, but the experiment is aimed at a single controller to realize ship motion control.

The experimental data are introduced into the simulation as a reference. Some studies use real environment and real trajectory data as references for controller verification. For example, when studying the influence of ship propeller dynamics and hydrodynamic characteristics on the ship’s position and speed control, [2] refers to the AIS data received from the Rotterdam Port Authority to scale the trajectory of an inland oil tanker during the two-hour voyage as a reference path to realize the simulation of trajectory tracking. AMBS-P algorithm is proposed to realize the research of autonomous docking and undocking of ships [3]. The construction of the simulation environment is based on the real geographical environment data, which makes the research as close as possible to the actual experimental environment. In order to test the feasibility and effectiveness of SHM-C&G&A prediction model and GCWOA, the measured heave and pitch displacement data of C11 container ship in irregular waves are used and verified in the virtual environment [4].

The research object and experimental results are given directly, and the testing platform is not described in detail. When studying the heading control problem of unmanned surface vehicle (USV) with uncertainty based on model free adaptive control (MFAC) theory, the experimental results are directly described by the “Dolphin IB” test platform [5]. Aiming at multi-ship collision avoidance in restricted waters, the effect of deep reinforcement learning is verified by model tests of three self-propelled ships. The experiment focuses on the design of the communication process between multiple ships, and there is no description of the testing platform [6]. [7] presents an automatic docking system of a single propeller and single rudder ship using a path tracking algorithm. A shipboard control system was built to perform automatic docking control. This paper introduces the basic situation of the experiment and the surrounding environment and does not elaborate on the platform in detail.

The intelligent ship and the construction of the testing platform are introduced, in which only the hardware control system or software system is described in detail. For example, [8] makes an experimental evaluation on the autonomous navigation and collision avoidance of the maneuvering intelligent guided ship. The hardware structure of the ship navigation and control platform is presented in detail. The software architecture is a series of programs developed under LabVIEW and MATLAB, but the detailed design interface is not shown. To solve the problem of small ship modeling and trajectory tracking control of the unmanned surface vehicle (USV) in time-varying navigation environments and speed, an adaptive PF controller based on a characteristic model and full coefficient adaptive control algorithm is designed and implemented. The paper introduces the structure and composition of the research object and the basic hardware, without elaborating on the software system [9]. A hybrid prediction model of ship motion and attitude based on long short memory (LSTM) neural network and Gaussian process regression (GPR) is proposed by [10]. To verify the effectiveness of the prediction model proposed in this paper, experiments under static and moving states are designed respectively. The experiment focuses on the inertial measurement unit (IMU) which collects the attitude information of the ship’s continuous motion and its acquisition process. For the study of ship docking and undocking, its planning and control mode is divided into two stages. This paper introduces the external sensors equipped with the experimental ship and the main demo board. The software interface is not proposed in this paper [5]. [11] uses the real AIS data records of Port Houston, focusing on proposing a tangible visual analytic tool to provide explicit risk identification related to navigation safety, so it does not involve the description of the hardware system.

The intelligent ship and the construction of the testing platform are introduced in detail, including software and hardware systems, but the experiment is aimed at a single controller to realize ship motion control. For example, [12] proposes a new design method of global controller, which can make the underactuated ship move along a smooth path under environmental disturbance. This paper briefly introduces the design structure of the model ship and the software interface written in Labwindows/CVI and verifies the effectiveness of this controller through experiments. [13] introduces an autonomous surface ship motion planning, guidance and control system in the actual marine environment. This paper introduces the hardware configuration and measurement accuracy of the whole system. The software architecture is mainly programmed by LabVIEW. The guidance system is based on a line of sight (LOS) navigation controller and does not involve other controllers. By developing a self-propelled autonomous system based on Wifi for a ship model with an Internet of things (IoT) function to provide a solution, maneuvering and navigation tests can be performed in an indoor environment without any complex mechanical structure. This paper introduces the software and hardware design of the whole system in detail, and the software is written based on LabVIEW. This system mainly realizes the research of fluid dynamics, not autonomous motion control. [14].

To sum up, due to the different emphases of the article, the description of the testing platform is simplified to varying degrees. In these studies, different hardware systems, software systems and controllers are designed according to the required functions. Among them, the software system is mostly realized by LabVIEW. In addition, the purpose of designing the testing platform in the study is to verify the effect of a single controller and make corresponding improvements. However, the improvement of a single controller is limited, and cannot meet the needs of a variety of scenarios. In this case, it is necessary to consider using a variety of controller-switching methods to adapt to the current navigation scene of the ship. Based on the above research and analysis, this paper builds a ship testing platform, designs a control board according to the functional requirements and carries the corresponding hardware equipment. The functions of this platform include testing the model characteristics of different model ships, perception, decision-making, and controller effects in different scenarios. This paper only shows the test function of controllers. The software display system is written based on Python language and can realize the motion control of the ship based on a single type controller or optimized switching motion controllers. In this paper, the “optimal switching strategy of motion controller” is designed and implemented given the large turning error of a single type of controller. The software system is scalable and the types of controllers can be added. The contributions of this paper are as follows:

  • The structure and design process of the ship testing platform is described.

  • The design idea and research process of “optimal switching strategy of motion controller” is summarized.

  • The control mode based on the combination of traditional controller and intelligent optimization controller is designed.

  • A new method is proposed to reduce the large turning error of ships during navigation.

This paper is organized as follows. Section 2 describes the framework of intelligent ship and its motion control system. Section 3 introduces the design of an intelligent ship motion controller and the optimal switching strategy. Section 4 gives the testing platform system design. Section 5 shows the actual experimental process. Section 6 summarizes the research results and gives future research prospects.

Section snippets

The framework of intelligent ship and its motion control system

The architecture of intelligent ships can be divided into functions like intelligent hull, intelligent navigation, intelligent facilities, intelligent energy efficiency and others. The main structural relationship is shown in Fig. 1.

Ship intelligent navigation system is an integrated system integrating the functions of perception, planning and decision-making, control and execution, and remote/shore-based monitoring systems. The functions of each system of typical intelligent navigation system

Design of intelligent ship motion controller and optimal switching strategy

Based on the structure of the ship’s intelligent navigation system, this section introduces the design of the controller. This paper designs multiple controllers switching to enhance the ship’s adaptability for various types of motion control.

Design of ship testing platform

The ship test platform can be divided into a hardware system and a software system. The hardware system includes a model ship, driving equipment, a control cabinet and some sensing equipment. The software system consists of a control interface and underlying functional modules written based on Python language. It is implemented on a computer configured with a Core i7-10750 CPU, 2.6 GHz with 16 GB RAM.

Experimental verification of optimal switching strategy for intelligent ship control

Based on the testing platform, this paper first realizes the ship heading control. Through the experimental results, the advantages and disadvantages of the controller are compared and analyzed, and the appropriate motion controller is set to optimize the switching strategy. The experiments were carried out in the environment shown in Fig. 11.

Conclusions and future work

This paper designs a ship testing platform, and takes the traditional controller and intelligent optimization controller as an example to solve the problem of the large ship turning error. Firstly, by analyzing the framework of the intelligent ship and its motion control system, the intelligent navigation system of the model ship in this paper is displayed. Then the optimal switching strategy of the ship motion controller is proposed for a variety of special scenarios in ship navigation. In

CRediT authorship contribution statement

Le Wang: Conceptualization, Methodology, Software, Writing – original draft. Shijie Li: Formal analysis, Software, Writing – review & editing. Jialun Liu: Project administration, Formal analysis, Funding acquisition, Writing – review & editing. Yuanchao Hu: Hardware and Software design. Qing Wu: Funding acquisition, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (17)

There are more references available in the full text version of this article.

Cited by (6)

Supported by Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), China (SML2021SP101), National Natural Science Foundation of China (62003250).

View full text