智能船舶远程驾驶负荷调控算法及仿真数据

智能船舶远程驾驶控制中心的人员任务分配与负荷调控是保障航行安全的关键问题。本数据集基于人员绩效模型（HPM）和近端策略优化（PPO）强化学习算法，构建了智能船舶远程驾驶负荷调控系统，并生成了完整的训练和仿真数据。数据集包含三个主要部分：（1）算法代码子集，含HPM模型、分层神经网络架构、任务分配系统等5个Python模块；（2）模型参数子集，包括超参数配置和训练损失记录；（3）仿真数据子集，记录系统决策、模型输出、组别统计和9名操作员的详细状态，共54个数据字段。仿真数据覆盖3个轮班时段，24小时的完整运行过程。本数据集可为智能船舶人因工程、强化学习在海事领域的应用、多智能体任务分配等研究提供算法参考和数据支撑。Personnel task allocation and workload regulation for the remote piloting control center of intelligent vessels are pivotal challenges for ensuring navigation safety. This dataset develops a remote piloting workload regulation system for intelligent vessels based on the Human Performance Model (HPM) and Proximal Policy Optimization (PPO) reinforcement learning algorithm, and generates complete training and simulation data. The dataset comprises three core components: (1) Algorithm Code Subset, which includes 5 Python modules such as the HPM, hierarchical neural network architecture, and task allocation system; (2) Model Parameter Subset, covering hyperparameter configurations and training loss records; (3) Simulation Data Subset, which records system decisions, model outputs, group statistics, and detailed statuses of 9 operators, with a total of 54 data fields. The simulation data covers the full 24-hour operational process across 3 shift periods. This dataset can provide algorithmic references and data support for research in areas including human factors engineering of intelligent vessels, applications of reinforcement learning in the maritime domain, and multi-agent task allocation.

搜集汇总

背景与挑战

背景概述

该数据集基于人员绩效模型和PPO强化学习算法，构建了一个智能船舶远程驾驶负荷调控系统，并提供了相关的算法代码、模型参数和仿真数据。仿真数据覆盖24小时运行过程，包含多操作员状态和决策记录，可用于智能船舶人因工程和强化学习应用研究。