Design of an Experimental Framework for Testing and Evaluation of Autonomous Ships Based on the Turing Test

In recent years, there has been a significant growth in intelligent ship technology. As the automation, intelligence, and autonomous development of ships continue to advance, it becomes crucial to conduct extensive testing and verification to ensure the rationality, reliability, and integrity of their functions and performance. Developing a mature and efficient intelligent testing and evaluation system becomes a prerequisite for further advancement in intelligent ship technology. This paper proposes a method for designing an experimental framework for intelligent navigation testing of ships, utilizing the Turing test. The first step involves building the necessary hardware and software experimental platform based on the requirements of driving behavior. Additionally, standardized experimental steps are designed to align with the characteristics of the platform. Subsequently, an experimental data evaluation method is proposed, based on probabilistic statistics and considering different navigation scenarios and characteristics of navigation data. This method relies on the principles of the normal distribution law. Finally, a human-machine confrontation experimental framework is constructed, based on the Turing test principle. The effectiveness of this framework is verified by integrating the experimental platform and experimental parameters. The proposed experimental framework serves as a reference and provides theoretical support for intelligent ship navigation testing, with the aim of promoting the overall development of intelligent ships.