EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
2411. Proposal for an ML-supported ETCS L2 Data Point Plan Review Process
Invited abstract in session WD-45: Planning Techniques for Decision Support, stream Decision Support Systems.
Wednesday, 14:30-16:00Room: 30 (building: 324)
Authors (first author is the speaker)
| 1. | Salome Vogel
|
| Department of Civil and Environmental Engineering, Institute of Railway Engineering | |
| 2. | Arturo Crespo Materna
|
| Department of Civil and Environmental Engineering, Technical University of Darmstadt, Institute of Railway Engineering | |
| 3. | Cedric Steinbach
|
| Department of Civil and Environmental Engineering, Technical University of Darmstadt, Institute of Railway Engineering |
Abstract
The timely rollout of ETCS L2 as a standardized train control system in Europe is depending on capacities of human planning experts, which are especially limited in the field of plan review.
Central requirements for ETCS L2 data point (DP) plan review are, to assess the completeness and correctness of the ETCS L2 DP plan, following a different logic than the plan creation.
This proposal presents an approach to review the completeness of the ETCS L2 DP plan. The rule-based nature of the underlying planning rules and regulations leads to the choice of a supervised learning (SL) algorithm. To assess completeness, in this approach the ETCS L2 DP plan under review is segmented into functional and spatial areas, with specific DP types assigned to particular reference points within those areas. The SL algorithm learns in which cases, which type(s) of DP are required within each area. Training data are existing ETCS L2 DP plans for different scenarios by human experts. An important benefit of the SL algorithm is that it learns from pattern in the training data instead of applying the planning rules and regulations 1:1, hence allows a plan review following a different logic than plan creation. Additionally, the plan review supported by the proposed SL algorithm provides decision support for plan reviews carried out by human experts. Thereby it facilitates a faster rollout of ETCS L2, contributing to interoperability, resilience and thus increasing rail capacity in Europe.
Keywords
- Decision Support Systems
- Machine Learning
- Railway Applications
Status: accepted
Back to the list of papers