Xie, Xinran (Beijing Jiaotong University), Liu, Jing (Beijing Jiaotong University), Huang, Yuxuan (Beijing Jiaotong University), Pei, Xuan (Beijing Jiaotong University), Hou, Taogang (Beijing Jiaotong University)

RailTopoSim: An Automatic Scene Generation and Simulation Framework for Rail Infrastructure Based on Topological Ontology

Scheduled for presentation during the Invited Session "S19a-Artificial Transportation Systems and Simulation" (TH-LM-T19), Thursday, November 20, 2025, 12:10−12:30, Surfers Paradise 1

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 18, 2025

Abstract

Accurate simulation and perception testing are essential for developing intelligent railway systems. However, generating realistic scenes with complex topology and sensor data remains challenging due to limited access to infrastructure and the high cost of multi-sensor deployment. We propose RailTopoSim, a topology-driven framework for automatic railway scene generation and simulation. It incorporates a rail-specific ontology to model spatial and logical relationships among key infrastructure components, including tracks, stations, signals, and switches. A procedural engine maps these elements into 3D environments using Unreal Engine 4, supporting dynamic, high-fidelity scenario construction. The system also includes sensor simulation modules that produce labeled multi-modal datasets (e.g., images, point clouds) by simulating train movement and virtual sensors. Experiments show that RailTopoSim efficiently generates diverse, topologically accurate scenes, reducing generation time by 93.7% compared to manual workflows. The framework provides a reusable, scalable solution for perception algorithm validation and data-driven research in railway automation.