Distributed Train Control Synchronization Using Virtual ML Models

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current distributed train control systems lack the capability to provide real-time, synchronized data for accurate prediction and mitigation of synchronization problems between distributed computer control systems, limiting the ability to generate a virtual representation of train position, configuration, and operational status, as well as perform predictive failure analysis.

Innovation Solution

A machine learning system with a data acquisition hub, analytics server, and client terminal that models actual train control systems using virtual system models, monitors real-time configuration data, and generates warnings for synchronization deviations, enabling adjustments to synchronize distributed computer control systems within a threshold deviation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If distributed computer control systems are used for train operation, then operational flexibility and remote control capability are improved, but synchronization accuracy and data consistency deteriorate

Engineering Contradiction:

Improveoperational flexibilityVSAvoidsynchronization accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The system continuously monitors configuration data from multiple distributed control systems and compares it against expected values. When deviations are detected, the system generates alerts and enables corrective actions to restore synchronization, creating a closed-loop feedback mechanism that maintains data consistency across distributed systems.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

A centralized control system acts as an intermediary between distributed control units. This intermediary collects configuration data from all distributed systems, performs synchronization analysis, and coordinates corrections across the network, mediating between the autonomy of individual control units and the need for global consistency.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If real-time monitoring of configuration data is implemented, then synchronization problems are detected earlier, but system complexity and computational requirements increase

Engineering Contradiction:

Improvesynchronization reliabilityVSAvoidsystem complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system creates virtual copies of configuration data from distributed control systems and maintains these copies in a centralized repository. By working with these copies rather than directly manipulating the original distributed data, the system can perform monitoring and analysis operations without adding significant complexity to the underlying distributed architecture.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The system pre-establishes expected configuration values and synchronization thresholds before operation begins. During runtime, it only needs to compare incoming data against these pre-computed references, avoiding the need for complex real-time analysis algorithms and reducing computational requirements.

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If machine learning models are used for predictive failure analysis, then prediction accuracy is improved, but data processing time and computational resources increase

Engineering Contradiction:

Improveprediction accuracyVSAvoiddata processing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system applies machine learning models selectively to only those configuration parameters and control systems that show signs of potential synchronization issues. Rather than continuously applying complex ML algorithms to all data, the system intervenes partially and targetedly, reducing overall processing time while maintaining prediction accuracy for critical issues.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS11787453B2Maintenance of distributed train control systems using machine learning

Publication Date: 2023.10.17 PROGRESS RAIL SERVICES CORP

AI summary

A machine learning system for maintaining distributed computer control systems for a train may include a data acquisition hub communicatively connected to a plurality of sensors configured to acquire real-time configuration data from one or more of the computer control systems. The machine learning system may also include an analytics server communicatively connected to the data acquisition hub. The analytics server may include a virtual system modeling engine configured to model an actual train control system comprising the distributed computer control systems, a virtual system model database configured to store one or more virtual system models of the distributed computer control systems, wherein each of the one or more virtual system models includes preset configuration settings for the distributed computer control systems, and a machine learning engine configured to monitor the real-time configuration data and the preset configuration settings. The machine learning engine may warn when there is a difference between the real-time configuration data and the preset configuration settings, the difference being indicative of at least two of the distributed computer control systems being out of synchronization by more than a threshold deviation.