Research on run-time risk evaluation method based on operating scenario data for autonomous train

Ru Niu
Sifan You

Document Type

Journal Article

Publication Date

2022

Subject Area

mode - rail, place - asia, place - urban, planning - methods, planning - safety/accidents, technology - intelligent transport systems

Keywords

Run-time Risk Evaluation, Autonomous Train, Operational Design Domain (ODD), Dynamic Bayesian Network (DBN), Fuzzy Set Theory

Abstract

Recent years witness the focus of the research of next-generation railways on risk situation awareness and safety decision-making to enhance the autonomy of unmanned trains. However, complex environmental factors make it difficult to assess the risks of train operation. Thus, it is of great necessity to clearly monitor the scenario parameters under which the train control system is designed to work, and to infer real-time risk through the collected scenario data. This paper first clarifies the key scenario parameters that need to be collected during the operation according to the concept of Operational Design Domain (ODD) and operating scenario. The key parameters and their dependencies are used to derive the Dynamic Bayesian Network (DBN) structure. Second, for data probability uncertainty, Fuzzy Set Theory is introduced, within the framework of which a fuzzy dynamic reasoning process is presented by monitoring the scenario data deviation. Finally, a case of real-time risk evaluation and analysis of the accident of Singapore MTR is explicated to demonstrate its contribution to operating data-based runtime risk analysis.

Rights

Permission to publish the abstract has been given by Elsevier, copyright remains with them.

Comments

Accident Analysis and Prevention Home Page:

http://www.sciencedirect.com/science/journal/00014575

Link to Full Text

Share

COinS