智能船舶远程驾驶行为知识库及远程驾驶负荷量化实验数据集

智能船舶远程驾驶任务的负荷量化是保障远程操控安全性的基础性工作。本数据集包含智能船舶远程驾驶行为知识库与远程驾驶负荷量化实验数据两部分核心内容。其一，行为知识库采用认知任务分析方法，系统性地分解了避碰、离靠泊、双船轮巡监控、故障处理及团队协作等5种典型驾驶场景的任务序列，并融合Petri网与VACP多资源模型建立了驾驶负荷评定框架。其二，远程驾驶负荷量化实验数据是通过招募持证船员，在新一代航运系统高仿真度模拟平台上，系统性开展了涵盖不同难度的远程驾驶模拟实验所采集。数据内容全面，包括船舶操纵与航行轨迹数据，以及同步采集的心率、呼吸、心电、皮电等多通道高频生理原始信号，并进一步提取了心率变异性（HRV）的时域与频域特征。数据采集与处理方面，采用专业生理记录仪与模拟器内嵌功能实现，并通过严格的预处理流程（包括滤波、去伪影）对原始信号进行加工。质量保证方面，通过事件标记技术确保了行为、生理等多源数据的精确时间对齐，并实施了贯穿全程的质量控制，保障了数据集的可靠性与一致性。本数据集可为智能船舶人因工程、远程驾驶任务分析以及驾驶负荷评估技术等领域的研究提供关键数据支撑，尤其有助于推动基于多源信息融合的驾驶员状态客观评估模型的开发，并为相关系统的优化设计与安全标准制定提供科学依据。

Load quantification of remote driving tasks for intelligent ships is a fundamental work to ensure the safety of remote control. This dataset consists of two core components: the knowledge base of intelligent ship remote driving behaviors and the experimental data for quantifying remote driving loads. First, the behavior knowledge base uses cognitive task analysis to systematically decompose the task sequences of five typical driving scenarios, including collision avoidance, berthing and unberthing, dual-ship patrol monitoring, fault handling, and team collaboration. It also integrates Petri nets and the VACP multi-resource model to establish a driving load assessment framework. Second, the experimental data for quantifying remote driving loads were collected by recruiting certified crew members to conduct systematic remote driving simulation experiments of varying difficulty on a high-fidelity simulation platform for the new-generation shipping system. The data content is comprehensive, including ship maneuvering and navigation trajectory data, as well as synchronously collected multi-channel high-frequency raw physiological signals such as heart rate, respiration, electrocardiogram (ECG), and electrodermal activity (EDA). Additionally, time-domain and frequency-domain features of heart rate variability (HRV) were further extracted. For data collection and processing, professional physiological recorders and simulator-embedded functions were used, and raw signals were processed through strict preprocessing procedures (including filtering and artifact removal). In terms of quality assurance, event marking technology was used to ensure precise temporal alignment of multi-source data such as behavioral and physiological data, and full-process quality control was implemented to guarantee the reliability and consistency of the dataset. This dataset can provide key data support for research in fields such as human factors engineering of intelligent ships, remote driving task analysis, and driving load assessment technology. It is particularly helpful for promoting the development of objective driver state assessment models based on multi-source information fusion and providing scientific basis for the optimal design of related systems and the formulation of safety standards.

搜集汇总

背景与挑战

背景概述

该数据集包含智能船舶远程驾驶行为知识库与远程驾驶负荷量化实验数据两部分核心内容，涵盖5种典型驾驶场景的任务序列和多通道高频生理原始信号。数据集通过专业设备采集并经过严格预处理，为智能船舶人因工程和远程驾驶任务分析提供关键数据支撑。