基于基元组合和语义约束的虚拟高速铁路环境智能建模方法研究

Simulation of general coupling system of high-speed railway is a highly complex dynamic process. Different specialized objects have complex and dynamic relations, which are uncertainty and randomness. Thus, it is very important that these simulation data are implemented analysis, reorganization and visualization in a virtual geographical environment. Because it can enable the abstract and complicated phenomena to be more actual and intuitive, and then facilitates decision-making support, such as exploration, explanation, prediction and planning. This project aims to establish an intelligence modelling method of virtual high-speed railway scene, which can effectively adapt the dynamic complex simulation environment. Based on the unification semantic understanding and the knowledge mapping, the intelligent system can separate the domain knowledge from the modelling operation. So the complexity and the difficulty can be efficiently reduce to improve the automaticity of modeling virtual high-speed railway environment. After analyzing the spatial characteristic and the relational attribute of various domains object in the high-speed railway system, we will firstly research on high-speed railway ontology integration and reasoning based on spatial semantic relationships. Meantime the primitive ontology knowledge library is also designed. Then the multi-level semantic restraint model of primitive combining is constructed. Finally we use the multi-agent system technology to explore intelligent guidance and control method in modelling process of virtual high-speed railway environment and to construct an intelligent modelling system. The expected research output will be directly used to the scientific simulation experiment and the management analysis of digital high-speed railway. It is also expected to offer a new idea and theoretical basis for further study and the application of modeling in complex virtual geographic environment.

高速铁路大系统耦合动力学仿真是一个高度复杂的动态过程，各专业对象之间存在复杂动态的联系，具有不确定性和随机性，因此，在虚拟地理环境中实现仿真数据的分析、重组与表示，并进行知识发现、模拟、预测和规划决策就具有明显优势。本课题旨在建立一套适应于动态复杂仿真情形下的虚拟高速铁路环境智能建模理论与方法，使领域知识和建模操作得以分开，降低多领域协作建模的复杂度与难度。先通过深入研究高速铁路各领域对象的空间特征和关系属性，建立基于空间语义关系的高速铁路环境模型基元本体；然后剖析场景建模过程中各对象之间的交互操作特征，建立虚拟高速铁路环境基元组合的多层次语义约束模型；进而探索动态复杂虚拟环境的多因素特征及相互制约关系，建立基于多智能体的虚拟高速铁路环境自动建模方法。研究成果可直接用于高速铁路数字化仿真的科学实验和管理分析，对我国高铁核心技术的自主创新和高铁系统的运营安全具有重大意义。

高速铁路大系统耦合动力学仿真是一个高度复杂的动态过程，各专业对象之间存在复杂动态的联系，具有不确定性和随机性。因此，在虚拟地理环境中实现仿真数据的分析、重组与表示，并进行知识发现、模拟、预测和规划决策就具有明显优势。本课题建立了一套适应于动态复杂仿真情形下的虚拟高速铁路环境智能建模理论与方法，使领域知识和建模操作得以分开，降低了多领域协作建模的复杂度与难度。主要研究成果如下：. ①构建了高速铁路基元模型库及其关系表达图，实现了基元模型的准确定义和清晰描述，可更好地支持标准化建模操作和高效管理分析；. ②通过建立多层次空间语义约束规则，提出了规则引导下的复杂场景三维建模方法，解决了传统建模操作复杂、繁琐、规范性差等问题；. ③提出了知识模板构建及其重用方法，解决了多类型场景建模知识不能有效重用的问题；. ④提出了场景映射与实例化方法，实现建模输出与场景渲染解耦，解决三维场景的绘制引擎单一固化问题。. 研究成果可直接用于高速铁路数字化仿真的科学实验和管理分析，对我国高铁核心技术的自主创新和高铁系统的运营安全具有重大意义。