Research on the Methods Evaluating the Performance of the Autonomous Berthing System of Intelligent Ships

Xiangyu Zhang,Yingjun Zhang

DOI: https://doi.org/10.1109/imcec55388.2022.10020064

2022-01-01

Abstract:With the development of autonomous berthing technology, it is very important to evaluate the autonomous berthing system of intelligent ships scientifically and rigorously if it is to be used commercially. However, there are few types of research on the evaluation for the performance of the autonomous berthing system of intelligent ships, nor is there is any reasonable evaluation model at present. Given this problem, considering the particularity of an intelligent system, according to the experience and consultation from many experts in the field, 13 indicators are introduced in this paper to establish an evaluation index system for the performance of the autonomous berthing system of intelligent ships from four aspects, namely safety, economy, efficiency and intelligence. The evaluation of performance of the autonomous berthing system of intelligent ships is a multi-stage and multi-index evaluation process, and there are interactive relations among the indexes. DEMATEL (Decision Making Trial and Evaluation Laboratory) and ANP (Analysis Network Process) method are used in the field of evaluation for the performance of the autonomous berthing system of intelligent ships. The interactive relationship among 13 indicators is established through the DEMATEL method, and the causal relationship diagram have been drawn. The weight of each performance indicator is determined through the ANP method. An evaluation model is established based on the improved GRA (Grey Relational Analysis) method. According to the characteristics of each indicator, the corresponding membership function is established for each indicator, and all indicators are converted into benefit-oriented indicators. In the data obtained through the membership function, not only the dimensionalization of original data is eliminated, but the discreteness of the data also increases, which avoids the tedious calculation of standardized data in GRA meanwhile providing a more scientific and reasonable basis for judgment. Finally, an intelligent container ship is taken as an example. Through the data collected from the actual ship experiment, the performance scores and final scores of this container ship have been obtained through the berthing process, which verifies the accuracy of the model and provides a reference for the study on the evaluation for the performance of the autonomous berthing system of intelligent ships.