Research & Advanced Engineering Compositional Variant Management and its Application in Embedded Software Development ProSTEP ivip Symposium 2010, Berlin 2010-04-29 Dr. Markus Kees, Ford Forschungszentrum Aachen GmbH Achim Seibertz, PROSTEP IMP GmbH PROSTEP IMP GmbH and Ford Forschungszentrum Aachen GmbH 2010 Dr. Markus Kees,
PROSTEP IMP Integrated Mechatronic Processes Facts Established in October 2005 8 offices in Germany Methods & Systems for Product-& Portfolio Management Process Consulting for MechatronicsSystem Development Specialization Solution provider for integrated processes, methods and technologies for the development of mechatronic products Methods, Systems & Integrations for Requirements Engineering & Management Methods, Systems & Integrations for Compositional Variant Management Projects > 20 successful customer projects in the automotive industry 3 successfully finalized research projects 1 active research project 3 ProSTEP ivip Association projects Methods & Systems for Model BasedFunction Development & Orientation Methods, Systems & Integrations for Wiring Harness Development
Challenges of Variant Management The number of theoretical possible Vehicle Variants is higher than the number of sold Cars Product Variants Trend Complexity Driver Increase of Product Variants (Market Segmentation) Increase of Variant Product Functions Increase of Variant Engineering Data Engineering Data Functions Current Situation Variants are specified and managed in decentralized, discipline-specific Data Management Systems and Development Tools Decentralized & Tool specific Systems of Rules are used for Configuration Goal Development of a pervasive Variant Modeling & Management Methodology Support of the entire Product Creation Process Compatible with existing and established Specification Methods
Product Variants correlates with Variants in the Engineering Data along the Development Process very, very, very simplified representation
Product Variants correlates with Variants in the Engineering Data along the Development Process very, very, very simplified representation
Product Variants correlates with Variants in the Engineering Data along the Development Process very, very, very simplified representation
Product Variants correlates with Variants in the Engineering Data along the Development Process very, very, very simplified representation
Product Variants correlates with Variants in the Engineering Data along the Development Process very, very, very simplified representation
E/E- Engineering Data are stored in different, non-pervasive IT-Systems Variants are defined discipline specific Goal: Achieve Traceability & Process-Support by System Integration Goal: Reduce Development Effort and increase Quality by Tool- Integration very, very, very simplified representation
Variant specification in a PDM-System (Example: Siemens PLM Teamcenter) Specification of Variant Positions Specification of the corresponding Conditions Configuration of an Instance
Variant Specification in Requirements Management Systems (Example: IBM Rational DOORS) 150% Specification Assignments to Configuration are expressed by Attributes Configuration by means of Filtering for T 100% Specification
Raumkonzept Raumangebot vorne inkl. Ein- / Ausstieg Raumangebot hinten inkl. Ein- / Ausstieg Ablagekonzept Kofferraum / Zuladung VDA Volumen min/max in l, Länge, Breite und Tiefe des Kofferraums in mm Gesamtfahrzeug Zugänglichkeit Kofferraum Gesamtfahrzeug Höhe der Ladekante in mm, Höhe der Ladeöffnung und Breite der Ladeöffnung in mm Heckscheibenneigung in Grad Gesamtfahrzeug The Idea of compositional Variant Management in the PDP Customer Req. Functional Spec. System Arch. Software Arch. Implementation < Variant Data and Structures along the PDP S 1 S 2 S 3 S 4 S 5 Transformation into a generic Variants Representation Unified Representation of different Variant Information Composition of the different Variants Models by means of the overlaying Product Variants model Analysis Analysis of Sub- and Total- Model (e.g. Consistence) SAT CSP FOF OPT
Methodology of Compositional Variant Management (Example: E/E-Development) 1 Rule Based Specification of Product Variants by Feature- Models 3 2 Definition of the Functional Architecture & Structure incl. Variants 4 Linking (Product Variants with Function Variants) Linking (Function Variants with Engineering Data) 5 Derive Configuration of Engineering Data from configured Product Model Analysis Engineering Data (Example E/E) Requirements Architecture Implementation Test
Example of a simple Feature-Model - Product Variants of a Vehicle Car mandatory optional Chassis Transmission Engine Cruise Ctrl alternative or Off-Road Sport Automatic Manual Electric Gasoline Diesel Standard Adative needs excludes V1 V2 With feature models complex variant product families and data can be described in a user friendly graphical manner Furthermore the benefit of feature models is the transformability into propositional logic, which can be analyzed to check for consistence, dead features, etc.
Methodology of Compositional Variant Management (Example: E/E-Development) 1 Rule Based Specification of Product Variants by Feature- Models 3 2 Definition of the Functional Architecture & Structure incl. Variants 4 Linking (Product Variants with Function Variants) Linking (Function Variants with Engineering Data) 5 Derive Configuration of Engineering Data from configured Product Model Analysis Engineering Data (Example E/E) Requirements Architecture Implementation Test
Research & Advanced Engineering Application @ Research & Advanced Engineering - Europe Compositional Variant Management in Model Based Embedded Software Development PROSTEP IMP GmbH and Ford Forschungszentrum Aachen GmbH 2010 Dr. Markus Kees,
System landscape: Today Research & Advanced Engineering Process data Requirements SW Architecture Modeling & Code Generation Verification & Validation Documentation Calibration Development System DOORS Matlab/Simulink Office CalDesk TargetLink Embedded- Validator & Tester MS Word, VISIO, RCP/HIL Data Management System easee.edm The tool chain of model based Software and function development has been enhanced since 1995 (Series Code Generation in 1999, formal Verification in 2004) Dependency of requirements and fulfillment of these had been ensured through a thorough process integration All process files have been managed with the easee.edm Ford Forschungszentrum Aachen GmbH 2010 Dr. Markus Kees 10/05/10
System landscape: Potential Future scenario Research & Advanced Engineering Process data Requirements Architecture Modeling & Code Generation Verification & Validation Calibration Development System easee.rqm AUTOSAR Architecture Tools Matlab/Simulink TargetLink Embedded- Validator & Tester CalDesk RCP/HIL Data Management System Documentation Generated easee.sdm easee.edm easee.tdm easee.cdm Compositional Variant information easee Engineering Data Backbone Ford Forschungszentrum Aachen GmbH 2010 Dr. Markus Kees 10/05/10
Linking of Engineering Data Research & Advanced Engineering Linking of engineering data of any kind. Ford Forschungszentrum Aachen GmbH 2010 Dr. Markus Kees 10/05/10
Methodology of Compositional Variant Management - Variant Management at FFA Research & Advanced Engineering 1 Feature Model 2 4 Configuration Model Parameter Model Parameter Sets Offline variant management tool with meta data exchange to backbone and link to development tool Representation and configuration of feature model model configuration parameters Ford Forschungszentrum Aachen GmbH 2010 Dr. Markus Kees 10/05/10
Methodology of Compositional Variant Management - Simulink Integration Research & Advanced Engineering Development System XML Active/selected system descriptions VM-Data Data Dictionary System descriptions VM-Data Global data lists XML Data Management System easee Engineering Data Backbone Meta data exchange File exchange Ford Forschungszentrum Aachen GmbH 2010 Dr. Markus Kees 10/05/10
Development 2010 Research & Advanced Engineering Process data Requirements Architecture Modeling & Code Generation Verification & Validation Calibration Development System easee.rqm Matlab/Simulink CalDesk AUTOSAR Architecture Tools TargetLink Embedded- Validator & Tester RCP/HIL Data Management System Generated Documentation easee.sdm easee.edm easee.tdm easee.cdm easee Engineering Data Backbone Ford Forschungszentrum Aachen GmbH 2010 Dr. Markus Kees 10/05/10
Open challenges and Future Vision Research & Advanced Engineering Integration pilots of RQM and SDM Methodical integration of VM GUI into easee landscape Graphical specification & representation of all variant information analogue to Feature Modelling Rule based configuration Completion of future target TDM and CDM Full traceability of all objects transparent to the user Analysis capabilities of variant data Ford Forschungszentrum Aachen GmbH 2010 Dr. Markus Kees 10/05/10
Research & Advanced Engineering Outlook PROSTEP IMP GmbH and Ford Forschungszentrum Aachen GmbH 2010 Dr. Markus Kees,
Raumkonzept Raumangebot vorne inkl. Ein- / Ausstieg Raumangebot hinten inkl. Ein- / Ausstieg Ablagekonzept Kofferraum / Zuladung VDA Volumen min/max in l, Länge, Breite und Tiefe des Kofferraums in mm Gesamtfahrzeug Zugänglichkeit Kofferraum Gesamtfahrzeug Höhe der Ladekante in mm, Höhe der Ladeöffnung und Breite der Ladeöffnung in mm Heckscheibenneigung in Grad Gesamtfahrzeug The Idea of compositional Variant Management in the PDP & its realization Customer Req. Functional Spec. System Arch. Software Arch. Implementation < Variant Data and Structures along the PDP S 1 S 2 S 3 S 4 S 5 Transformation into a generic Variants Representation Unified Representation of different Variant Information v.control Analysis Composition of the different Variants Models by means of the overlaying Product Variants model Analysis of Sub- and Total- Model (e.g. Consistence) SAT CSP FOF OPT
PROSTEP IMP Concept Tool v.control Product line and variant specification using feature modeling Domain specific import or input of variant development artifacts Compositional linkage of product variants and development artifacts Rule based configuration to compose consistent productand development data Attribute based optimization of product lines Powered by Fraunhofer ISST
Research & Advanced Engineering Thanks for your attention! Dr. Markus Kees Hybrid Vehicle Technologies / Electronics & Control Alternative Powertrains Ford Research & Advanced Engineering Europe e-mail: mkees@ford.com Achim Seibertz Head of Methods & Technology PROSTEP IMP GmbH e-mail: achim.seibertz@prostep.com PROSTEP IMP GmbH and Ford Forschungszentrum Aachen GmbH 2010 Dr. Markus Kees,