Train Control LanguageTeaching Computers Interlocking

By: J. Endresen, E. Carlson, T. Moen1, K. J. Alme, Haugen, G. K. Olsen & A. Svendsen

Synthesizing Software ModelsGenerate Train Station Models Automatically

Presenter: Rahma Al MahruqiCISC 836

By: Andreas Svendsen, Øystein Haugen, and Birger Møller-Pedersen

Train Control LanguageTeaching Computers Interlocking

Outline

• Motivation

• Objectives

• Computer Based Interlocking CBI

• Domain Specific Modeling Language DSL

• Syntax of TCL

• TCL Semantics

• Advantages of TCL

Motivation

Bridging the gap between trains and computers

with a specially designed language that enables

the signaling experts to create consistent train

interlocking systems.

Objectives

• ABB and SINTEF organizations have developed TCL for specifying

and automating the train interlocking systems.

• Combine the station drawing with the train routes and interlocking

tables in the TCL.

• Generate automatically the programmable logic controller PLC code

and test cases from the models created in the TCL language.

• Describes DSL for specifying stations and to allow automatic

generation of interlocking tables and the code for the Computer

Based Interlocking CBI

Terminologies

• DSL : Domain Specific Language

• DSM Domain Specific Modeling

• TCL : Train Control Language

• PLC: Program Language Controller

• CBI : Computer Based Interlocking

• EMF: Eclipse Modeling Framework

• GMF : Graphical Modeling Framework

Introduction

• Interlocking system controls the basic elements of the train station

such as signals, switches, truck circuits and allocate train routes to

avoid collisions.

• Interlocking system ensures that the route is safe by reading the

status of the elements in the route (e.g. tracks, switches, signals) to

see if they comply with the logic of the interlocking system.

• Prevents dangerous train movement on a train station by

giving a “clear” signal to a train only if the requested

route is safe.

Interlocking System

Interlocking System

• Development of interlocking source code is a time-consuming

process requiring a large amount of resources.

• The logic is illustrated by an interlocking table, and realized by the

interlocking source code, in the form of functional blocks of code

that are executed by the PLCs in their control of the station.

CBI Development Workflow

Domain Specific Language

• DSL is a modeling language dedicated to a particular problem

domain.

• Bridge between the experts of an application domain and the

experts of computer science.

Train Control Language

• DSL done through a metamodel, which automates the production of

source code for computers controlling train stations

• Improves the completeness and consistency of the specifications

and reduce the human effort.

• Applies UML class diagrams to define metamodels.

Metamodel

• Model of the language which defines the language concepts

and the relations between these concepts.

• Characteristics of a Metamodel:

- Composition (Stations contain TrackCircuits etc.),

- Attributes (Station has a name) and

- Relations (TrainRoute refers to a number of

TrackCircuits).

A small piece of the TCL Metamodel

Track Inheritance Hierarchy.

Concrete Syntax of TCL

• Open source Eclipse and its modeling framework EMF

technology used to define TCL.

• Graphical Modeling Framework GMF used to define the

graphics of the terms of the metamodel.

• GMF technology helps produce an editor for TCL based on the

metamodel and the graphic definitions.

• The editor is embedded into the eclipse environment.

The Semantics of TCL

Advantages of TCL

• Provides a clear interface between the realms of computer science

and signaling.

• Makes the signaling experts capable of creating advanced computer

systems without knowing the details of computer programming.

• Provides automatic generation of several different artifacts

previously manually developed.

• Much of the tedious and error-prone manual controls can be

eliminated and the highly competent signaling experts can instead

be used to refine the designs through the TCL language.

Conclusion• TCL can improve the development process of interlocking systems.

• Signaling experts will be able to design the systems directly.

• The interlocking systems will be generated directly from the graphic

representation which is produced for validation and tests.

• The traditional manual work will be eliminated or minimized.

Synthesizing Software ModelsGenerate Train Station Models Automatically

Motivation

• Presents an approach for automatic synthesis of software models

which are increasingly being used for representing software

applications at a high abstraction level.

• Model analysis is useful for performing model-based testing which

involves generating test-cases for the application.

• Software applications can be produced quicker and more reliable by

automating the task of building models

• A number of software models can be created to validate and verify

a code generator.

Objectives

• Description of an approach for synthesizing software models

automatically based on a formal definition of the DSL and user-

defined properties.

• Illustrate real example from the train domain TCL, where train

station models are generated automatically.

Terminologies

• DSL is defined by an abstract syntax, a concrete syntax and

well-defined semantics.

• Accompanied by code-generators and frameworks.

• TCL is a graphical DSL for modeling train stations and

generate configuration code for a framework which

controls train station signaling systems.

• Alloy is a formal language used to find and produce

intended DSL models.

TCL Metamodel

TCL concrete syntax

TCL Model Components

• Repository model, containing the abstract syntax,

• Diagram model storing the location of each element on the canvas.

• Includes algorithms for calculating the diagram model based on the

repository model.

• This allows the graphical diagram model to be reconstructed if the

repository model is exposed to changes.

TCL Graphical Editor

Alloy

• Light-weight declarative language using relational calculus for

modeling a system formally .

• Formal validation and verification of a system using mathematical

notation to express the system precisely, to prove its correctness.

• Supplied with an analyzer tool based on first-order-logic performs

automatic analysis of Alloy models.

How Alloy works

• Alloy Analyzer only performs analysis within a user-specified

scope, restricting the number of elements of each type in the

model

• When the Alloy Analyzer searches for a solution, it populates

all signatures in the model with model elements up to the

user-specified scope.

• An Alloy model typically consists of signatures, fields, facts,

predicates and assertions.

Model Synthesis

Generate models automatically based on specifications, for:

• Reducing development time of models by generating a basis which

can be extended.

• Modeling large sets of models to reduce time-consuming.

• Generate a model that is close to the target application model.

Synthesizing TCL Model

• Generating TCL Models

• Adding Dynamic Properties

• Optimizing Alloy Models

Synthesizing TCL Model

Generating TCL Models

TCL Generated Stations- Static Properties

TCL with Dynamic Properties

• Dynamic properties is defined as a quality that can only be determined by dynamic analysis, such as model simulation.

• Using dynamic analysis, certain properties of the model can be checked or simulated before the model is generate, eg. The number of tracks of the station can give an indication of how many trains the station can handle simultaneously.

Optimizing Alloy Models

• The Alloy Analyzer performs complete analysis within the user-specified

scope. Complete analysis involves populating all possible models and

discarding the ones where the constraints (facts) do not hold. Therefore,

• When the scope increases, the number of possible combinations to form

solution models, and thus the analysis time, increases exponentially.

Conclusion

• Generate application models automatically based on a well defined

specification of a DSL.

• Use Alloy to define the semantics of the DSL and illustrated the

approach using TCL.

• Alloy is used because of its uniform notation and its automatic

analyzer.

• Using Alloy as a specification language provides the possibility to

specify the operational semantics of DSL .

Conclusion

• Performing model synthesis reduce the cost of developing

applications, and several applications can be generated to test a

code generator.

• Thorough knowledge of Alloy as specification language and TCL is

required to apply the proposed approach.

Critique

Advantages:

• Well organized paper

• Lot of figures

• Good references

Limitations:

• Approach with limit to 3 track station only- generate large set of

station using dynamic approach

• Perform case studies and implementing the approach with other DSL