Railway Traffic Conflict Detection via a State Transition Prediction Approach

Abstract

Conflict detection and resolution is one of the most important tasks in daily railway traffic management, although it is still difficult to solve all its aspects. In fact, the aspect of conflict detection has not been amply studied. In this paper, an approach of traffic state prediction and conflict detection, based on proper state transition maps (STMaps) and corresponding relation matrices, is proposed. First, the traffic state sequences, which mainly concern infrastructure status and train movement information, are studied. These state sequences are expressed as segment and route state vectors and kept in corresponding state-domain tables (SDTables). The empirical state transitions are then applied to detect irregular states in a dynamic traffic environment. Furthermore, the structural constraints of infrastructure topology and route compatibilities are represented in matrices to aid the calculation and prediction of potential conflicting situations. Scenarios such as train delay and infrastructure failure are designed to test the proposed approach. The test results show that irregular states can be efficiently detected and potential conflicts can be further identified, and the detailed conflict information is also approachable.