Mise en page 1
Transcription
Mise en page 1
01 AUTO:Mise en page 1 12/05/10 17:27 Page 12 Automotive & Transports Working Group Gilles LE CALVEZ, WG President [email protected] VALEO 12 “Our strategy is based on a threefold objective: boosting technological innovation in order to increase industrial performance and enhance the attractiveness of Paris region. This strategy fully leverages the contribution to international Standards. To date, 28 projects labelled by Systematic were initiated by the Automotive & Transports Working Group. This Working Group has demonstrated its ability to develop innovative projects structured by roadmaps and, most importantly, to capitalize on their results by integrating them into a platform of economic valorization. Our Working Group developed at European level a coordination of several clusters of competitiveness in France, Germany, Spain and Austria to be able to promote new technologies in standard such as Autosar, Cenelec or ISO 26262". 01 AUTO:Mise en page 1 12/05/10 17:27 Page 13 EMBEDDED SYSTEMS AND ELECTRONICS, AT THE HEART OF THE TRANSPORTATION SECTOR With a GDP share of 15%, the transport industry (aviation, rail and automotive) structures very strongly France’s economic and industrial activities. Paris Region is the first French region in this sector with 156 000 employees representing 19% of national employment. Regarding R&D, the automotive industry, as an example, employs in the Ile-de-France 17 500 people including 6 600 researchers, representing 75% of french researchers in this field, 67% of the research staff in this industry and 72% of expenditure on R&D. THE AUTOMOTIVE & TRANSPORTS WORKING GROUP WITHIN SYSTEMATIC This Working Group aims at mastering the complexity of electronic architectures and embedded-systems to meet the need for technological development concerning embedded systems in automotive, railway and truck industries, leveraging the expertise and competences of all these sectors. ◗ R&D Financed Projects: 28 ◗ Partners: 71 including : ◗ 22 SMEs ◗ 27 Large companies ◗ 22 Research institutes and universities ◗ Total Investment: 168.4 M2 The Automotive & Transports Working Group is structured around six technology areas, called “Development Axes”: ◗ ◗ ◗ ◗ ◗ ◗ Electronics Architecture Dependability of monitoring systems Software tools and Methods Diagnosis Human-machine interfaces Algorithms & data fusion for localisation and vision It also cooperates with the Telecom Working Group Xavier APOLINARSKI, Michaël FOURNIER, WG Vice-President [email protected] CEA Representative of Permanent Secretariat [email protected] SYSTEMATIC Steering Committee Members ALSTOM TRANSPORT ARMINES PARIS ARMINES PARIS CEA SACLAY CEA SACLAY CNRS PARIS CONTINENTAL AUTOMOTIVE France CONTINENTAL AUTOMOTIVE France DELPHI FREESCALE SEMICONDUCTOR FREESCALE SEMICONDUCTOR INRETS LYON INRETS VERSAILLES INRIA ROCQUENCOURT INRIA ROCQUENCOURT INTEMPORA RENAULT GUYANCOURT RENAULT GUYANCOURT RENAULT TRUCKS RENAULT TRUCKS UNIVERSITE DE TECHNOLOGIE DE COMPIEGNE VALEO BOBIGNY VISTEON VISTEON Didier VAN DEN ABEELE Claude LAURGEAU Sébastien LEROY Riadh CAMMOUN Jean-Marc ALEXANDRE Véronique VIGUIE DONZEAU-GOUGE Serge BOVERIE Louis-Claude VRIGNAUD Kamel MAAZIZ Denis GRIOT Gérard MANIEZ Guy BOURGEOIS Jean-Marc BLOSSEVILLE Anne-Céline LAMBALLE Laurent KOTT Gilles MICHEL Olivier GUETTA Christian BALLE Claude COVO Bernard FAVRE Ali CHARARA Philippe GOUGEON Jacques JAUBER Jean-Paul GULIA [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] [email protected] Automotive & Transports WG 13 01 AUTO:Mise en page 1 12/05/10 17:27 Page 14 Diagnosis 0-DEFECT Outil de Diagnostic Embarqué de Faisceaux AUTomobiles Embedded systems are equiped with diagnosis functions, but that of the wired network, which is a critical system, is not considered. In 2006 the project SEEDS aimed at studying the feasibility of a wire diagnosis system usable in garage for maintenance or embedded for the follow-up of the harness. The objectives of 0-DEFECT are to study and implement methods more adapted to embedded. The current method gives a ten centimeters precision on the localization of a hard defect, but is not compatible with embedded constraints (EMC and interaction of the diagnosis signals with those of the network). Other methods will be implemented, the models will be refined and more powerful prototypes will be built and tested in the vehicle, aiming at proving viability and feasibility for an embedded system. ON GOING PROJECT CONTACT Marc OLIVAS CEA LIST +33 (0)1 69 08 48 83 [email protected] PARTNERS TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS The main scientific and technical objectives in the project are: Large companies: DELPHI, FREESCALE, PCA SA, VOLVO TRUCKS Research institutes, universities: CEA LIST, INRIA, SUPELEC LGEP ◗ online embedded diagnosis of real topology automobile cable network (in a concurrent way to the normal functioning of the network), ◗ detection and localization of hard defects with a precision of a few centimeters, at low cost, PROJECT DATA Coordinator: CEA LIST ◗ solving the localization ambiguity inherent to ramified networks, ◗ study of the detection of the degradation premises of connectors, follow-up of the evolution of parameters representative of the degradation of a connector and way of measuring them, ◗ study of the embedded diagnosis architecture, ◗ microelectronic integration study, in an objective of diffusion in the automotive domain shortly after end of the project. Call: ANR Start date: August 2009 Duration: 36 months Global budget (M2): 4.1 STATUS - MAIN PROJECT OUTCOMES SP1 (specifications) provided some results: ◗ an analysis of intermittent failures showed that 25% were due to cable defects, ◗ specifications for the diagnosis chip were provided, ◗ the integration study of the chip is ongoing and showed the best solution is to use a diagnosis chip connected to the existing architecture of the calculator. In the meantime, the partners have begun to write the consortium agreement, on the basis of those already signed, for example in the SEEDS project. Automotive & Transports WG 14 Funding (M2): 1.9 01 AUTO:Mise en page 1 12/05/10 17:27 Page 15 Electronics Architecture ARAMIS ON GOING PROJECT The ARAMIS project is a part of Num@tec Automotive. The project ojectives are the following: ◗ Set-up a "Best practices" development center between SMEs and Scientific laboratories in order to increase the AUTOSAR Software Components development efficiency. The main objectives is to promote the use of Model-Based design methodology associated to automatic embedded code generation and to enhance the validation means. ◗ Create a model library optimized for embedded code generation on Cooler Systems. ◗ Initiate a continuously improved design flow between SMEs and laboratory by measuring CONTACT Sébastien DUBE GEENSYS +33 (0)2 53 46 20 56 [email protected] and improving the productivity. ◗ Finalize the project by a long-term contract for a common laboratory based on Scientific studies used on concrete industrial uses cases. PARTNERS TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS ◗ PathCrawler technology (CEA): Used to generate automatically tests to cover all cri- SMEs: GEENSYS, SHERPA Research institutes, universities: CEA-LIST teria on software execution (example: All branches, statements… in the source code). This tool generates tests cases to execute all the branches of the executable code. This tool asset that tests oracles are known and defined. ◗ Agatha Technology(CEA): Used to generate automatically tests to cover testing objectives given in the models. Then, the models (simulink models for example) can be tested before the embedded C source code generation on the simulation engine. ◗ AUTOSAR code generation technology (GEENSYS): Used to generate automatically AUTOSAR C embedded source code from the Matlab/Simulink models which describe the expected behaviour of the software. The Software C source code complies with AUTOSAR Standard and allow execution on AUTOSAR target incluing the AUTOSAR BSw Stacks. The set of these technologies allows to perform a significant increase of productivity and reliability in automotive embedded software development. STATUS - MAIN PROJECT OUTCOMES The ARAMIS technologies have been successfully deployed by GEENSYS on air conditionning system model provided by SHERPA. Simulation and embedded code on target execution can be compared with the tool chain. PROJECT DATA Coordinator: GEENSYS Call: FEDER0 Start date: January 2009 Duration: 16 months Global budget (M2): 0.8 Funding (M2): 0.4 Related Systematic project(s): EDONA The CEA-LIST technologies have been also deployed in order to generate execution tests and coverage measurement. Automotive & Transports WG 15 01 AUTO:Mise en page 1 12/05/10 17:27 Page 16 Algorithms & data fusion for localisation and vision Automotive Robust Operating Services ON GOING PROJECT Built on top of the RTMaps concept, AROS is the next generation of Real-Time Component-Based Design platform: ◗ AROS is Agile: Designed by and for Automotive Engineers. Eases application development and maintenance. ◗ AROS is Distributed: AROS applications can be transparently distributed on a heterogeneous network. CONTACT Arnaud DE LA FORTELLE ARMINES +33 (0)1 40 51 92 55 [email protected] ◗ AROS is Dynamic: Components can be connected / disconnected dynamically. ◗ AROS is Reliable, Safe & Secure: Total failure confinement and security of data exchange through AES. AROS: a tool to master the most innovative automotive projects... TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS AROS provides to the automotive engineer a set of innovative tools or libraries: ◗ DOHC (Distributed Objects for Higher Level C) eases the development of robust applications in C, while providing a unique real-time memory management system. PARTNERS Large companies: VALEO SMEs: INTEMPORA SA Research institutes, universities: INRIA, MINES PARISTECH ◗ Cables (the communication library) introduces a new subscribe/push protocol and eases the development of client/server applications (including web 2.0 applications) while providing state-of-the-art compression and encryption of communications. ◗ Chassis (the I/O library) provides a way to develop applications or components independently of the underlying platform: Windows, UNIX or embedded real-time system. ◗ The new execution engine, based on the optimistic execution paradigm - a concept used in state-of-the-art parallel simulation systems -, is able to provide dynamic components management, transparent distribution over cores or systems while ensuring realtime deterministic execution. Coordinator: ARMINES Call: ANR Start date: September 2008 All together, these tools enable the development of more complex automotive applications, involving complex communication systems among vehicles and/or infrastructure, distribution and redundancy of computation, redundancy and multiple sensor fusion (radar, GPS, cameras, etc.) Duration: 36 months Global budget (M2): 2.1 Funding (M2): 1 STATUS - MAIN PROJECT OUTCOMES All the major components of AROS have been implemented (execution and communication engines). Two demonstrations using AROS are being designed: a Stop'n'Go Prototype by VALEO, and a Robust & Distributed Traffic Management System by INRIA. The challenge is to release a competitive COTS product by december 2010. Automotive & Transports WG 16 PROJECT DATA 01 AUTO:Mise en page 1 12/05/10 17:27 Page 17 Software tools and Methods Cost-efficient methods and processes for safety relevant embedded systems The project is adressing the design of safety-critical embedded systems in different domains: aerospace, railways, industrial automation and automotive. These domains are facing increasing complexity and variety of products, coupled with increasing regulatory constraints while costs, performances and time to market are constantly challenged. Potential cross-domain synergies in the area of methods and tools need to be exploited. CESAR targets significant and conclusive innovations in two cross-domain topics: Requirements engineering and Component based engineering. Gathering a critical mass of key partners, CESAR defines a european tool platform for safety critical systems design and development, with strong industrial support. This tool platform will bring to the market CESAR innovations, and will be open to further innovations from other contributors in the research community on embedded systems. TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS The overall technological innovation is an integrated environment for the design and development of safety relevant embedded systems. This environment (RTP, Reference Tool Platform) embodies standard languages, methods, and tools supported by European tool vendors. The technical challenges relate to Requirements Engineering and Component Based Development, including in particular formalization of multi viewpoint, multi criteria and multi level requirements, applied to design space exploration comprising multi-view, multi-criteria and multi level architecture trade-offs. Multi-viewpoint based development processes targets early verification and validation of non-functional properties. The objective is to assure that not only functional aspects but also safety, costs, robustness, timeliness, etc. are captured and documented in a form allowing validation and verification to be performed. Multi-criteria based design processes allow for optimization of designs to multiple objectives functions for such extra-functional characteristics of designs, a key for achieving competitive products. Multi-level design flows cover all stages from initial concepts through requirements through specification to design and implementation, typically crossing multiple stages in supplier chains. STATUS - MAIN PROJECT OUTCOMES Relying on use-cases and scenarios from Aerospace, Automotive, Automation and Railway, CESAR is strongly industry driven. Benefiting from this multi-domains point of views, CESAR addresses safety aspects of transportation and other societal mobility and environmental demands. This will be a significant step forward in terms of industrial performance improvement that will help to establish de-facto standards and contribute to the standardization effort from a European perspective. ON GOING PROJECT CONTACT Joseph MACHROUH THALES RESEARCH & TECHNOLOGY +33 (0)1 69 41 57 21 [email protected] PARTNERS Large companies: ABB, ACCIONA, AIRBUS, ALENIASIA, ASF, ASTRIUM, AVL, CENTRO RICERCHE FIAT, DASSAULT SYSTEMES, DANIELI AUTOMATION, DELPHI, EADS, ELSAG DATAMAT, ESI-TECNALIA, HELLENIC AEROSPACE INDUSTRY, HISPANO-SUIZA, INFINEON TECHNOLOGIES, MESSIER-BUGATTI AND TURBOMECA, SAGEM, SIEMENS, THALES, VOLVO SMEs: ABSINT, CRITICAL SOFTWARE, ESTEREL TECHNOLOGIES, FORMAL SOFTWARE CONSTRUCTION, GEENSYS, OSC, THE VIRTUAL VEHICLE COMPETENCE CENTER Research institutes, universities: ARISTOTLE UNIVERSITY OF THESSALONIKI, CEA, CNRS, DLR, FRAUNHOFER, INDUSTRIAL SYSTEMS INSTITUTE, INRIA, KUNGLIGA TEKNISKA HÖGSKOLAN, NORWEGIAN UNIVERSITY OF SCIENCE AND TECHNOLOGY, NATIONAL TECHNICAL UNIVERSITY OF ATHENS, OFFIS, ONERA, OXFORD UNIVERSITY, SINTEF, THE UNIVERSITY OF MANCHESTER, UNIVERSITÀ DI BOLOGNA UNIBO, UNITS PROJECT DATA Coordinator: AVL / GERHART GRIESSNING Call: ARTEMIS Start date: March 2009 Duration: 36 months Global budget (M2): 58 Funding (M2): 27 Related Systematic project(s): FLEX-EWARE, GENCOD, LAMBDA, MOVIDA, RT-SIMEX, USINE LOGICIELLE, VERDE Automotive & Transports WG 17 01 AUTO:Mise en page 1 12/05/10 17:27 Page 18 Diagnosis DIAFORE Diagnostic de Fonctions Réparties COMPLETED PROJECT ◗ Online diagnostics to improve the safety and maintenance of trucks: detect, locate and automatically identify sensor failures in trucks during operation. ◗ The manifestation of a failure on one or more electronic sensors can cause unexpected and dangerous behaviours of the truck for the safety of the driver and his environment. Furthermore, the prolonged detention of a truck due to a failure results in financial penalties for the company. The increasing complexity of electronics in trucks makes more difficult the identification of the faulty sensors. Fault diagnosis of electronics in garage (offline) exists today. The automatic diagnosis of faults during operation of the truck (online) is still a challenge. ◗ The project objective is to propose solutions for future electronic design by including at the same time the means of diagnosis. CONTACT PROGRESS BEYOND THE STATE OF THE ART PARTNERS From modelling of breakdowns to integration of diagnostics in embedded electronics: the working method in the project followed the V's cycle. A first step involved modelling breakdowns in the electronic sensors of the truck, described in a rigorous language, non verbose (e.g. Simulink, Formal language). In this representation, the electronic functions of the truck are described and the physical medium is abstracted (ECU, bus, sensors). The second step involved the development and integration of diagnostic features in all. The latter raises an indicator output when the failure is detected on an input function under diagnosis. Another output provides a code indicating the type of failure (e.g. failure of the speed sensor of the third rear wheel). The third step was the programming of these new features (software) in the electronic controllers of the truck with the constraint to make this integration as transparent as possible in the existing functions. At each step, the correct functioning of the whole was verified using simulation and testing. An additional study focused specifically on the modifications in the original functions of the truck to improve the efficiency of diagnosis (diagnosability analysis). MAJOR PROJECT OUTCOMES Large companies: RENAULT TRUCKS SMEs: SERMA INGENIERIE Research institutes, universities: CEA LIST, UPS11-LRI, UTC-HEUDIASYC PROJECT DATA ◗ Publications: Communications: • A Probabilistic Analysis of Diagnosability in Discrete Event Systems, Grèce, juillet 2008. • Optimal observability for diagnosability, Australie, septembre 2008. • Optimizing the system observability level for diagnosability,Grèce, octobre 2008. • Diagnosability of Input Output Symbolic Transition Systems,Portugal, septembre 2009. • Co-modelling and simulation with multilevel of granularity for real time electronic systems supervision. England, 1–3 April 2008. • Observability Checking to Enhance Diagnosis of Real Time Electronic Systems. October 27-29, 2008. Canada. • Symbolic Execution Based Model Checking of Open Systems with Unbounded Variables. TAP 2009, Switzerland. • Diagnosability verification with Petri net unfoldings, Chili, septembre 2009. • Smart Distance Keeping: modeling and perspectives for embedded diagnosis, Liverpool, Angleterre, 27-29 janvier 2010. Chapitre d'ouvrages: • [10] M. Khlif, M. Shawky. Book chapter: Enhancing Diagnosis Ability for Embedded Electronic Systems Using Co-Modeling. International Joint Conferences on Computer, Information, and Systems Sciences, and Engineering (CIS2E 07). December 3 - 12, 2007. ◗ Experimentations: • [1] a prototype tool to analyze the diagnosability of an electronic system. The demonstration of this prototype is done on the case study (SDK Smart Distance Keeping) simulated. It can conclude on the diagnosis of failure ie, there exists at least one instance of the operating system for which the failure can not be discriminated; • [2] a co-real/simulated online diagnostic demonstrator applied to the case of the SDK. The diagnosers identify the fault using analytical redundancy. The case study consists of 4 ECUs and a single CAN bus. Three ECUs are simulated on a computer and one is real. The simulated CAN bus is connected to the real CAN bus via a specific interface. The diagnosers are deployed on the 4 ECUs. One of them has the role of supervisor and provides the final diagnosis (sensor fault). Validation tests show that the diagnosers identify systematically failures injected the diagnosis latency due to CAN communications remains quasi constant; • [3] a prototype tool for analyzing the observability of failures, once the electronic system integrated into the truck. The demonstration of this prototype is done on the SDK implemented in the previous demonstrator. Automotive & Transports WG 18 Olivier HERON CEA LIST +33 (0)1 69 08 68 89 [email protected] Coordinator: CEA LIST Call: ANR Start date: February 2007 Duration: 43 months Global budget (M2): 3.7 Funding (M2): 1.9 01 AUTO:Mise en page 1 12/05/10 17:27 Page 19 Diagnosis DIAPA DIagnostic Automobile Par Apprentissage ON GOING PROJECT The goal of DIAPA is to improve the diagnosis of vehicles electronic subsystems. A comprehensive analysis of log-files containing records of components operation and default codes will be done using advanced data mining and statistical learning techniques. The aim of this project is to automatically handle the log-files, increasing the on-line processing, and establishing better links between diagnosis and design models. TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS ◗ Currently, few solutions give satisfaction in terms of collecting and processing diag- CONTACT Charles-Henri GARIH DELPHI +33 (0)1 34 30 34 76 [email protected] Abdallah FAHED UTC-HEUDIASYC +33 (0)3 44 23 52 15 [email protected] nosis information carried out on electronic subsystems of the vehicles. The aim of the automotive industry is to improve the “after-sale service” and reduce the costs of Non Trouble Found. ◗ Develop generic diagnosis methods regarding the large variety of possible configurations of subsystems and their failures. PARTNERS ◗ Establishing better links between diagnosis and design models. Large companies: DELPHI, FREESCALE, PSA, SERMA INGENIERIE Research institutes, universities: CNRS-LAAS, UTC PROJECT DATA Coordinator: DELPHI Call: ANR Start date: January 2008 Duration: 36 months Global budget (M2): 3.1 STATUS - MAIN PROJECT OUTCOMES Funding (M2): 1.2 ◗ Give a more precise study of system specifications and Logfiles diagnosis. ◗ Diagnosis algorithms adapted to logfiles with Heterogeneous log-files input data. Automotive & Transports WG 19 01 AUTO:Mise en page 1 12/05/10 17:27 Page 20 Electronics Architecture EDONA Environnements de Développement Ouverts aux Normes de l'Automobile ◗ Build a common repository for consistent positioning and usage of software tools. ◗ Build an interoperability platform for tool exchange and seamless operation over the development cycle. ◗ Build solutions for dependability and real-time on the basis of both innovative and current technology: • Specification and requirements traceability; • Functional and temporal validation; • Design of real-time and deterministic solutions; • HMI design and validation under safety constraints. ON GOING PROJECT CONTACT François OUGIER RENAULT +33 (0)1 76 85 75 72 [email protected] PARTNERS TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS This project deals with numerous innovations: ◗ Typical technical deliverables: • East-ADL2 timing extensions; • East-ADL2 graphical modeler; • East-ADL2 to Autosar bridge; • East-ADL2 to PharOS bridge; • Simulink to PharOS code gene-ration; • Numerical precision analysis tool; • Automatic test generator and execution monitor for Matlab models; • Automotive HMI description language, with converters to and from commercial HMI Large companies: CONTINENTAL, DELPHI, JOHNSON CONTROLS, PSA PEUGEOT CITROËN, RENAULT, THALES, VISTEON SMEs: ESTEREL, GEENSYS, INTEMPORA, KNOWLEDGE INSIDE, OBEO, SHERPA Research institutes, universities: ARMINES, CEA-LIST, INRIA, POLYTECHNIQUE, SUPÉLEC, design tool formats; • Scade to Autosar code generation. ◗ Technical transfers to other entities: • Eclipse components for Autosar toolsets > Artop User Group; • Eclipse components for modeling toolsets > Sphinx project of the Eclipse Foundation. PROJECT DATA Coordinator: RENAULT Call: FUI3 Start date: September 2007 Duration: 36 months Global budget (M2): 16.2 Funding (M2): 6.3 STATUS - MAIN PROJECT OUTCOMES ◗ October 12th, 2007: Project kick off. ◗ June 4th, 2009: Project midlife plenary meeting. ◗ May 6th, 2010: Project last plenary meeting. Automotive & Transports WG 20 Related Systematic project(s): MEMVATEX, O4A PHASE II 01 AUTO:Mise en page 1 12/05/10 17:27 Page 21 Algorithms & data fusion for localisation and vision E’MOTIVE Environment Modeling for Perceptive Intelligent Vehicles The aims of the E'MOTIVE project are the development of a simulator of detection system, the coupling of this simulator in a bundle of simulation softwares in order to create a numerical simulation platform for the design and validation of Advanced Driver Assistance Systems (ADAS), and the validation of the benefits of this simulation platform. Regarding the first objective, the idea is to ensure the robustness and reliability of the detection system by creating virtual sensor datas. Regarding the second objective, the focus is on the design and validation of the ADAS on this validated detection system. TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS The deliverables will be a detection system simulator, a numerical simulation platform for the design and validation of ADAS which will contain a set of applicative softwares with extended functionalities and interfaced each others, the results of the applications and validations stage. ON GOING PROJECT CONTACT Anne CHAMBARD LMS IMAGINE +33 (0)1 34 52 17 72 [email protected] PARTNERS Large companies: VALEO Intermediate size enterprise: LMS IMAGINE, OKTAL SMEs: CIVITEC, INTEMPORA, SHERPA ENGINEERING Research institutes, universities: CNRS-LASMEA, INRETS, INSA ROUEN, INSTITUT D’OPTIQUE GRADUATE SCHOOL, LCPC, TELECOM BRETAGNE PROJECT DATA STATUS - MAIN PROJECT OUTCOMES The project has been started on October 1st, 2008, for three years. Work Package 2 (Specifications) is achieved, WP3 (Development of the detection system), WP4 (integration) and WP5 (validations) are on-going. Coordinator: LMS IMAGINE Co-label: MOV'EO Call: FUI6 Start date: October 2008 Duration: 36 months Global budget (M2): 4.8 Funding (M2): 2.9 Related Systematic project(s): CSDL Automotive & Transports WG 21 01 AUTO:Mise en page 1 12/05/10 17:27 Page 22 Human-Machine interfaces GEDRIVER Green & Efficient Driver ON GOING PROJECT Experimental works of the geDRIVER project (“Green and Efficient Driver”) aim at developing and validating the simulator for training to ecological driving according to 3 major topics within the “serious game” axis: ◗ To develop the models allowing analysing eco-driving and the follow-up of performances in a context of training using simulator, which latter logically follows training on real vehicle. CONTACT Anton MINKO OKTAL SA +33 (0)5 62 11 92 92 [email protected] ◗ To reinforce the interest of driving experience in order to to support the attractivity of training. For that, the project will draw from the experiments and the techniques already implemented in video games. ◗ To contribute to enrich knowledge in the field of validity of driving training simulators. PARTNERS TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS In the context of emergent requirements regarding vocational training (European directive 2003/59 CE) and growing eco-driving trend recorded over the past years, geDRIVER aims at exploring the feasibility of affordable solutions for automotive training (desktop simulators). Consequently, "Serious Game" approach coupled with existing simulation suite SCANeR is considered in the framework of geDRIVER project, according to the following axes: ◗ enriched gameplay (based on Large companies: OKTAL SA, RENAULT Intermediate size enterprise: OKTAL SA Associated partners: MACIF, RENAULT TRUCKS SMEs: KEY DRIVING COMPETENCES SA Research institutes, universities: ARTS - INSTITUT IMAGE LEAD-CNRS, LABORATOIRE LE2I real life situations such as home-to-work or taxi itinerary) allowing to involve the driver into the simulation, ◗ virtual guiding metaphors (both visual and audible) allowing to push the simulation as far as possible to the driving reality, ◗ data reading / analysis from a real vehicle. A special attention is put on the evaluation of project's result (behaviour-based measures will be taken in order to issue recommendations and to make simulator-based training as close as possible to the real car-based training; simulator's sickness analysis is a part of evaluation). STATUS - MAIN PROJECT OUTCOMES The project will produce and test prototypes/demonstrators allowing to: ◗ integrate video "game"-based approaches and paradigms into "serious" automotive simulation, ◗ integrate new interactive visual and auditive guiding techniques, ◗ get real vehicle data (via CAN bus). Demonstrations will be available for light vehicles' and trucks' drivers. Automotive & Transports WG 22 PROJECT DATA Coordinator: OKTAL SA Call: SERIOUS GAMING Start date: December 2009 Duration: 18 months Global budget (M2): 1.2 Funding (M2): 0.7 01 AUTO:Mise en page 1 12/05/10 17:27 Page 23 Software tools and Methods HECOSIM Heterogeneous Co-simulation and system hybrid Simulation The design of a mechatronic system, inherently heterogeneous, is made currently with tools adapted to each sub system (various modelers and simulators). Each sub system is validated separately, and the integration tests are performed at the end of the development cycle (physical mock up); problems are discovered late and generate high cost to be corrected. The objective of the HeCoSim project is to provide simulation tools through two approaches of co-simulation and global simulation based on available tools to: ◗ control sizing and the integration of automotive components of heterogeneous nature, ◗ optimize an electronic architecture through model simulation, ◗ validate a system by scenarios filtered and generated by the tools, ◗ validate the embedded code on its target in its physical environment. PROGRESS BEYOND THE STATE OF THE ART Problems treated and solutions sought: ◗ remove a major technical issue on asynchronous co-simulation via the tool RT-Builder; ◗ create co-simulation sockets between the tool RT-Builder and the tools: Statemate, MatLab/Simulink, SystemVision, UNISIM; ◗ have a good representativeness and interoperability of the simulation models of the microcontroller Star12X and of the processor core ARM7 on the open source UNISIM tool; ◗ extend the AGATHA semantics to the Statemate and Simulink models; ◗ specify a filter for AGATHA to generate relevant tests regarding industrial criteria: undesirable cases, deadlocks, MC/DC approach. Consequences at technical level: ◗ extension of the simulator RT-Builder to co-simulation; ◗ achievement of a complete and continuous tool chain with the core activity tools; ◗ validation of the embedded code with its physical environment; ◗ achievement of a heterogeneous simulation core for AGATHA and a filter for the MC/DC approach. COMPLETED PROJECT CONTACT Jean-François BISSON VALEO +33 (0)1 45 13 82 55 [email protected] PARTNERS Large companies: VALEO SMEs: GEENSYS Research institutes, universities: CEA LIST, UTC-HEUDIASYC PROJECT DATA Coordinator: VALEO Call: ANR Start date: December 2006 Duration: 39 months MAJOR PROJECT OUTCOMES Global budget (M2): 2.7 ◗ Publications: Communications (conference): • IAEC’2008: “Improving Model Based Design Quality and Safety for Mechatronic Systems via Co-simulation”. • Workshop System@tic 2008. • EMM’2009: “A Model Based Design Compliant Framework for Mechatronics Systems Development”. • Workshop System@tic 2009. Articles: • Co-modelling and simulation with multilevel of granularity for real time electronic systems supervision, England, 1-3 April 2008. • Observability Checking to Enhance Diagnosis of Real Time Electronic Systems Canada, October 27-29, 2008. • Validation of Electronic Architecture Properties Analysis Based on Automotive Platform, Italy, July 4-7, 2010. • Diagnosability of Embedded Architectures; Premise of analysis methodology. • Symbolic Execution Techniques Extended to Systems in ICSEA 2009. Funding (M2): 1.4 ◗ Job creation: Two hiring, one at CEA-LIST/LSL and another one at Geensys. Automotive & Transports WG 23 01 AUTO:Mise en page 1 12/05/10 17:27 Page 24 Algorithms & data fusion for localisation and vision ICADAC Improved Camera based Detection under Adverse Conditions The ICADAC project is dedicated to road safety improvement under adverse meteorological conditions. Being able to detect and quantify these adverse conditions of operation using only a video signal is a strong challenge for driver assistances systems (ADAS). The first step is to detect and quantify this signal degradation, and identify their causes. It lead to estimate a confidence index on the operation of an ADAS based on cameras and its self-diagnosis. Meanwhile, it is possible to adapt the sensor operation, to dynamically adjust its parameters or to improve signal quality. It is also a way to bring further new driver assistances systems such as automation of fog lamps, control of adaptive tail-lights or driver warning in case of inadequate behaviours. TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS This project deals with numerous innovations: ◗ On-line diagnosis of the camera operation. A camera integrated in a vehicle must have a self-diagnosis functionality. This project will be one of the first to propose a method to qualify the video signal. ON GOING PROJECT CONTACT Didier Aubert LCPC - INRETS / LIVIC +33 (0)1 40 43 29 18 [email protected] PARTNERS Large companies: AUDI, VALEO Research institutes, universities: ARMINES, INRETS, IOSB, LCPC ◗ Detection and characterisation of night fog. At night, the fog is mainly perceived through the halo phenomenon and the backscattered veil. Both of them are due to the scattering of the beams of the ego-vehicle or of the other vehicles. This project will thus deal with a new fog context to, finally, achieve a complete fog detection and characterisation system. Coordinator: LCPC - INRETS / LIVIC ◗ Detection and characterisation of rain by an embedded camera. Sensors already exist for this purpose. The idea is to replace them by a camera system to get only one sensor dealing with several functions. Co-label: MOV'EO Call: ANR ◗ Image improvement. Thanks to the image improvement, existing ADAS based on camera sensor with benefit from this project. Start date: July 2009 Duration: 36 months ◗ Demonstration of new applications. ◗ Development of new test benches Global budget (M2): 3.8 dedicated to fog and rain. STATUS - MAIN PROJECT OUTCOMES WP1 task (requirements and specifications) is finished. That means a list of potential applications was established, some of them were selected, camera module requirements and specification is done. Three documents are delivered. We also started to equip vehicles, to design a testbench for night fog characterisation, started to develop methods to detect and characterise fog and rain, as well as methods to restore images altered by day fog and rain. Automotive & Transports WG 24 PROJECT DATA Funding (M2): 1.8 01 AUTO:Mise en page 1 12/05/10 17:27 Page 25 Human-Machine interfaces Interface Homme-Système à 10 ans COMPLETED PROJECT The target of IHS10 project is to develop a new R&D tool that will be dedicated to the Human-System Interfaces (HSIs) for the Automobile. It consists in a Virtual Reality (VR) equipment, which will contribute to reduce development cycle time and cost of new vehicles and platforms. It will be used on the higher level of the two branches Requirements / Validation - of the V-cycle, enabling actors to work in a collaborative way, replacing physical by virtual prototyping and improving quality of results. The field that this tool will cover is: geometry, styling, ergonomics, controls, displays, functions and associated softwares. The sensorial channels of this new VR equipment are: 3D vision, tactile / haptic and acoustic, in an immersive environment of the User. PROGRESS BEYOND THE STATE OF THE ART ◗ To use the new generation of HD projectors and 3D televisions for a better definition and to make the simulator more transportable. ◗ To develop and standardise the software and the equipment to facilitate the exchange of information between car manufacturers and equipment suppliers. ◗ To develop the sensory touch. CONTACT Jean DAUVERGNE VISTEON +33 (0)6 08 00 44 07 [email protected] PARTNERS Large companies: PSA PEUGEOT CITROËN, RENAULT, VALEO, VISTEON SMEs: HAPTION, INTEMPORA Research institutes, universities: ARMINES, CEA-LIST PROJECT DATA Coordinator: VISTEON Call: FUI1 Start date: March 2007 Duration: 36 months MAJOR PROJECT OUTCOMES Global budget (M2): 5.5 ◗ Products: Design and build a generic simulator for Human System Interfaces (HSIs) and related interior equipments of a vehicle , based on Virtual Reality techniques, enabling design, tuning and evaluation of HSIs by the Development Engineers, characterized by: • Co-localization (Immersion of User “in" the 3D VR environment). • Multi-sensorial channels: Visual, Haptic / Tactile, Acoustic. • Multi-domains: Vehicle Interior Components (geometry, aspect, color, textures), controls, displays, Softwares. • Interactive: Model and Softwares changes and adaptations. • High definition. Funding (M2): 1.9 ◗ Experimentations: PSA demonstrator (basis Peugeot 308). Automotive & Transports WG 25 01 AUTO:Mise en page 1 12/05/10 17:27 Page 26 Electronics Architecture IMOFIS Ingénierie des MOdèles des FonctIons Sécuritaires ◗ The purpose of IMOFIS is to develop an environment that assists safety engineers to analyse and verify systems and to provide evidence that these systems meet safety requirements. ◗ IMOFIS environment will handle models of systems and subsystems, models of safety analyses and models of design and safety processes. Furthermore, IMOFIS environment will handle the semantic relationships between all these models in order to maintain a global and consistent view of them. ◗ IMOFIS environment will be, as much as possible, a general environment not devised to particular application domains, processes or techniques. Consequently, it will rest upon a collection of models of models (meta-models) which will abstract from particular design and safety processes as well from particular specification, analyses and verification techniques TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS Industrial safety requirements verifications are, today, achieved manually and are based on informal specifications. IMOFIS project addresses the formalisation of both the safety requirements and their verification. Today, model driven techniques allow system designers to specify more accurately the overall system. Indeed, SysML has been introduced in the Alstom Transport methodological framework and Renault studies its introduction in its own process. However, SysML does not address the dysfunctional view of the system, this particular aspect is mandatory for the definition of safety requirements and the realisation of safety analyses. Several research works has proposed model transformation to derive safety analyses (semi-automatically) from the system model by mapping some system design pattern to safety analyses elements. Here, we propose another approach by developping the model of models of the elements required by safety engineer. Consequently, the safety process would be include in the same model framework that the design process, furthermore, the identification of semantics relations between this two viewpoints allows to assist the safety verification activities by model verification mechanisms included in the model of models. STATUS - MAIN PROJECT OUTCOMES The meta model of the safety viewpoint have been developed. It extends SysML to safety modelling activities (formelly, "hazards analysis"). The formalisation of model verification is ongoing work, it includes the definition of semantics relations between the safety viewpoint and the functional viewpoint. The next and final step will be the tools development and the application of the IMOFIS framework on representative industrial use cases. ON GOING PROJECT CONTACT Fernando MEJIA ALSTOM TRANSPORT INFORMATION SOLUTIONS +33 (0)1 57 06 13 73 luis-fernando.mejia @transport.alstom.com PARTNERS Large companies: ALSTOM TRANSPORT INFORMATION SOLUTIONS, RENAULT SMEs: ATEGO, OBEO, SYSTEREL Research institutes, universities: CEA-LIST, UTC-HEUDIASYC PROJECT DATA Coordinator: ALSTOM TRANSPORT INFORMATION SOLUTIONS Call: ANR Start date: June 2008 Duration: 36 months Global budget (M2): 4 Funding (M2): 1.6 Automotive & Transports WG 26 01 AUTO:Mise en page 1 12/05/10 17:27 Page 27 Human-Machine interfaces Integrated Human Modelling and Simulation to support Human Error Risk Analysis of Partially Autonomous Driver Assistance Systems The target of the ISI-PADAS project is to develop an innovative methodology to support risk-based design and approval of Partially Autonomous Driver Assistance Systems (PADAS) focusing on the elimination and mitigation of driver errors by an integrated Driver-Vehicle-Environment (DVE) modelling approach. TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS Main part of the DVE modelling approach will be an effective and working simulation of driver behaviour, based on modelling driver behavior and cognitive processes as well as PADAS behaviour, included in computerised simulations that can be applied in early development stages to predict driver behaviour including driver errors to support decisions between system design alternatives and to test the need for specialized assistance systems. ON GOING PROJECT CONTACT Jean-Paul GULIA VISTEON SYSTEMES INTERIEURS +33 (0)1 79 41 72 29 [email protected] Dr. Rainer HEERS VISTEON DEUTSCHLAND GmbH +49-2273-595-0 [email protected] PARTNERS Large companies: CENTRO RICERCHE FIAT, VISTEON SMEs: KITE SOLUTIONS Research institutes, universities: CEA, CIDAUT, DLR, INRETS, OFFIS, SUPELEC, TU BRAUNSCHWEIG, UNIVERSITY OF MODÈNA STATUS - MAIN PROJECT OUTCOMES ◗ ◗ ◗ ◗ ◗ ◗ ◗ Definition of relevant traffic scenarios. Driving studies and driving simulators studies (driving without PADAS support). PROJECT DATA Coordinator: OFFIS Task representation language. Ontology of driving task. Implementation of time for accelerated / real-time testing. Scientific advisory committee established. Organised scientific conference: Human Modelling in Assisted Transportation (HMAT), 2010. Call: FP7 Start date: September 2008 In progress: Duration: 36 months ◗ Complete DVE modelisation. ◗ Driving simulators studies (driving with PADAS support). Global budget (M2): 5.5 Funding (M2): 1.8 Automotive & Transports WG 27 01 AUTO:Mise en page 1 12/05/10 17:27 Page 28 ICT & Transport itransports 2.0 ON GOING PROJECT The project itransports 2.0 aims to develop from itransports content which is the only comprehensive information on transport at the national level, collaborative content on two important themes of transport for the coming period: disturbance and development of supply "bike". ◗ The entire contents of the disruption of transport will be supplied from the information transport networks, but also from feedback from users on itransports well as through social networks. CONTACT Dominique CARRE MOVIKEN +33 (0)1 60 33 03 03 [email protected] ◗ A PC "traffic" will be tested: it reaches the heart of a global information system on transport implemented on this occasion, including a website and an application interface with major mobile market. PARTNERS TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS SMEs: HAPLOID, MOVIKEN ◗ Scientifically, this project will allow for a standardized approach to the phenomenon of Research institutes, universities: INRETS disruption of transport and will better formalize the effort for research in this area, which has remained until now the preserve of transport operators: research conducted so far are relatively small and their results are not disseminated ◗ Benefits in terms of service: Coordinator: MOVIKEN • Controlling the phenomenon of interference and informational support better public is a prerequisite for the continuation expected and desired development of public transport. • The project will contribute 2.0 itransports powerfully as the first comprehensive solution for information on the disruption of transport networks. Call: WEB2.0 Start date: December 2009 Duration: 18 months Global budget (M2): 0.5 STATUS - MAIN PROJECT OUTCOMES The expected benefits of the project itransports 2.0 are important in different ways: ◗ They will forward the information and disturbances, which darken the image so that transport development is expected and desired, ◗ They will make this content available to those involved companies, communities, media, ... ◗ They will underpin the advantage Moviken took on the transport information market with a solution that will be unique in Europe. Automotive & Transports WG 28 PROJECT DATA Funding (M2): 0.2 01 AUTO:Mise en page 1 12/05/10 17:27 Page 29 Algorithms & data fusion for localisation and vision Logiciels d'observation des vulnérables COMPLETED PROJECT LOVe proposes to contribute to road safety by focusing primarily on the safety of pedestrians. The objective is to achieve embedded software to observe vulnerables. These software must be reliable and safe and quickly implantable on the range. An industrial design process of software has been adopted to clearly specify the technical constraints (sensors, computers), contextual (scenarios considered) and methods for validation of algorithms. LOVe is rich of a large variety of approaches using data laser, mono and stereo-vision both for the detection, localization, recognition and tracking of pedestrians. These solutions can then be combined with various levels of data fusion algorithms to improve location accuracy and certainty. CONTACT Laurent TRASSOUDAINE UNIVERSITE DE CLERMONTFERRAND 2 +33 (0)6 78 55 31 03 laurent.trassoudaine @univ-bpclermont.fr PROGRESS BEYOND THE STATE OF THE ART The most significant advance concerns the visual target tracking. Tracking of LASER targets becomes also very powerful. Although information is very fragmented on a pedestrian, the quality of the recognition is generally good. Groups of pedestrians can be identified too. The stereo has also grown in terms of speed, quality of the disparity map, segmentation of the road surface and segmentation of objects. Pedestrian recognition has especially increased in the rapidity of methods and a better classifiers choice. The data fusion significantly increases the rate of correct detection while decreasing the rate of false alarms. Especially, the lidar/mono-vision combination has provided good results. MAJOR PROJECT OUTCOMES ◗ Patents: 2 deposits concerning pedestrians recognition. ◗ Publications: • Journals - 2 book chapters - 36 conferences. • Object-level fusion and confidence management in a multi-sensor pedestrian tracking system; 2009. • Combination of partially non-distinct consonant beliefs: the cautious-adaptive rule; July 2009. • Pedestrian Detection and Tracking in Urban Environment using a Multilayer Laserscanner; to appear in 2010. • Using stereo-vision to improve the reliability of obstacle detection systems. • Submitted to IEEE Transactions on Intelligent Transportation Systems. • Visual pedestrian recognition in weak classifier space using nonlinear parametric models; San Diego, USA, october 2008. • Pedestrian Accident Context and Technologic Development in LOVe; Versailles Satory October 7 & 8, 2008. • Centralized fusion based algorithm for fast people detection in dense environment; 12-17 May 2009, Kobe, Japan. • A Bayesian Classification of Pedestrians in Urban Areas: The Importance of the Data Preprocessing; Seoul, Corée du Sud, 20 - 22 Août, 2008, pp. 195-201. • Robust obstacles detection and tracking using disparity for car driving assistance; San Jose, USA, January 2010. PARTNERS Large companies: RENAULT, VALEO Research institutes, universities: ARMINES/CAOR, ARMINES/CMM, CEA/LIST, CNRS/UTC/HEUDIASYC, CNRS/UBP/LASMEA, CNRS/UPS/IEF, INRIA/EMOTION, INRIA/ICARE, INRIA/IMARA, LCPC/INRETS/LIVIC, UNIVERSITE DE CLERMONTFERRAND 2 PROJECT DATA Coordinator: UNIVERSITE DE CLERMONTFERRAND 2 Co-label: MOVEO - VIAMECA Call: ANR Start date: September 2006 Duration: 40 months Global budget (M2): 8.5 ◗ Business creation: 3 start-up: • Arcure: Software for pedestrian detection; • Robocortex: Software for visual tracking; • Civitec: road scenes simulation. Funding (M2): 3.8 Automotive & Transports WG 29 01 AUTO:Mise en page 1 12/05/10 17:27 Page 30 Electronics Architecture MASCOTTE MAitriSe et COnTrôle des Temps d'Exécution Build methods, techniques and tools to keep pace with execution times of real-time applications using static analysis, test on simulators and measures on the real target. In order to implement new services in cars, the project studies the impact on the determinism of hardware features in microcontrollers. It defines good practices to use these mechanisms. COMPLETED PROJECT CONTACT Pascal SAINRAT IRIT - UNIVERSITE PAUL SABATIER +33 (0)5 61 55 84 25 [email protected] PROJECT RESULTS ◗ Technologies: Simulation (Harmless), static analysis (Otawa), test (PathCrawler). ◗ Publications: • Damien Hardy, Isabelle Puaut - WCET analysis of multi-level noninclusive set-associative instruction caches - 29th IEEE Real-Time Systems Symposium, Barcelona, Spain, December 2008. • Rola Kassem, Mikaël Briday, Jean-Luc Béchennec, Guillaume Savaton, Yvon Trinquet - Simulator Generation Using an Automaton Based Pipeline Model for Timing Analysis - International Multiconference on Computer Science and Information Technology, October 2008, Poland. PARTNERS Large companies: FREESCALE, RENAULT SMEs: GEENSYS Research institutes, universities: IEF, IRCCYN, IRISA/INRIA, IRIT, LAAS • Roman Bourgade, Clément Ballabriga, Hugues Cassé, Christine Rochange, Pascal Sainrat - Accurate analysis of memory latencies for WCET estimation - International Conference on Real-Time and Network Systems (RTNS 2008), Rennes, October 2008. • Marianne de Michiel, Armelle Bonenfant, Hugues Cassé, Pascal Sainrat - Static loop bound analysis of C programs based on flow analysis and abstract interpretation IEEE International Conference on Embedded and Real-Time Computing Systems and Applications (RTCSA 2008), Kaohsiung, Taiwan, August 2008. • Damien Hardy, Isabelle Puaut - Predictable code and data paging for real-time systems - 20th Euromicro Conference on Real-Time Systems, Prague, Czech Republic, July 2008. ◗ Experimentations: Use of the technologies on an automotive application. PROJECT DATA Coordinator: IRIT Call: ANR - 2005 Start date: January 2006 Duration: 42 months Global budget (M2): 3.5 Funding (M2): 1.4 Automotive & Transports WG 30 01 AUTO:Mise en page 1 12/05/10 17:28 Page 31 Software tools and Methods MEMVATEX Modeling Methods for Validation & requirements Traçability Definition of a methodology for the modeling and the validation of requirements trough a design flow for embedded real-time systems. This design flow guaranty the full traceability and the validation of requirements from their initial request down to their implementation. A special focus is put on non-functionnal requirements, like timing requirements. Project WebSite : www.memvatex.org COMPLETED PROJECT CONTACT Arnaud ALBINET CONTINENTAL AUTOMOTIVE SAS +33 (0)5 61 19 73 30 [email protected] PROJECT RESULTS ◗ Technologies: • A metamodel for requirements modeling, validation and traceability. • An implementation of the Requirement Meta-model as a specific UML-profile in Artisan Studio toolsuite. • The implementation in Artisan Studio toolsuite of the MARTE profile. • A methodology under Eclipse-EPF for the expression and the modeling of requirements which integrates the standards EAST-ADL2, MARTE and RIF. PARTNERS Large companies: CONTINENTAL SMEs: SHERPA ENGINEERING Research institutes, universities: CEA LIST, INRIA, UTC ◗ Publications: • RTCSA 2007 Multiform Time in UML for Real-time Embedded Applications. • SIES 2007 A multiform time approach to real-time system modeling. Application to an automotive system. • ECMDA 07 Model-based methodology for requirements traceability in embedded systems. •MCSE07 A Requirement-based Methodology for Automotive Software Development • ERTS08 The MeMVaTEx methodology: from requirements to models in automotive application design. • ACESmb 08 From high-level modelling of time in MARTE to realtime scheduling analysis. • ISORC09 Marte CCSL to execute East-ADL Timing Requirements. ◗ Experimentations: A demonstrator of the methodology on an automotive case study based on ARTiSAN Studio toolsuite and Rectify. • Job creation: 4 CDD. PROJECT DATA Coordinator: CONTINENTAL Call: ANR - 2005 Start date: January 2005 Duration: 36 months Global budget (M2): 2.1 Funding (M2): 1 Automotive & Transports WG 31 01 AUTO:Mise en page 1 12/05/10 17:28 Page 32 Software tools and Methods From System Modeling to S/W running on the Vehicle Digital mock-up (DMU) has been the automotive industry best practice that for many years has enabled many engineers from different disciplines to collaborate in the virtual world to define future real products better. However, until now this has been static – i.e. it did not take much account of product behaviour. In particular, the dynamic behaviour of embedded systems and software was neither managed nor simulated in DMU. The purpose of MODELISAR is to introduce functional mock-up (FMU), a next generation of methods, standards and tools to support collaborative design, simulation and test of systems and embedded software. The objectives of MODELISAR are threefold, to: ◗ Enable concurrent design of embedded systems and software, leveraging state-ofthe-art technologies: • The open Modelica language for component-oriented systems modelling and simulation; and • The AUTOSAR standard for automotive embedded software. ◗ Define advanced runtime interoperability interfaces to enable open co-simulation between virtual product models, especially in Modelica, and the executable embedded software, with various configurations. ◗ Deliver a smooth, traceable and integrated process for embedded systems and software across the product life, based on Dassault Systèmes V6 Product Life Management (PLM). TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS ◗ Define a new standard: FMI (Functional CONTACT François BICHET DASSAULT SYSTEMES +33 (0)1 61 62 44 30 [email protected] Patrick CHOMBART DASSAULT SYSTEMES +33 (0)1 61 62 74 63 [email protected] PARTNERS Large companies: ALTRAN, AVI, DAIMLER, DASSAULT SYSTEMES, LMS, VERHAERT, VOLKSWAGEN, VOLVO TEC, SMEs: ATB, DAVID, DYNASIM, EXTESSY, GEENSYS, INSPIRE, QTRONIC, SIMPACK, TRIALOG, TRIPHASE, TWT, VERHAERT Research institutes, universities: ARMINES, ARSENAL RESEARCH, DLR, FRAUNHOFER IIS EAS, FRAUNHOFER FIRST, FRAUNHOFER SCAI, IFP, UNIVERSITY OF HALLE Mock-up Interface) to enable FMI based co-simulation between various modeling, simulation tools and generated embedded software. ◗ Demonstrate the support of FMI by leading market tools, especially for Modelica and AUTOSAR. ◗ Demonstrate seamless integration of the Modelica and AUTOSAR design & verification flows on top of PLM V6 infrastructure. PROJECT DATA Coordinator: DASSAULT SYSTEMES Call: ITEA2 STATUS - MAIN PROJECT OUTCOMES Already available 1st version of FMI specifications wich enables the model exchange. FMI co simulation aspects are being defined. Several tool editors of Modelisar consortium have implemented FMI prototypes to support generation or execution of FMI compatible models. All the modelisar automative Use Cases are designed with involved models and tools, and are ready to realize the FMI based Proofs of Concept. Automotive & Transports WG 32 ON GOING PROJECT Start date: July 2008 Duration: 36 months Global budget (M2): 28 Funding (M2): NA 01 AUTO:Mise en page 1 12/05/10 17:28 Page 33 Software tools and Methods O4A2 Open 4 Autosar Phase II ON GOING PROJECT Proposed a set of methods, tools and technologies that respect Autosar objectives and needs for automotive ECU: Safety with the new standard ISO2626-2, independence regarding hardware platform, ease application design and validation. TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS CONTACT Jean-Sébastien BERTHY GEENSYS +33 (0)2 53 46 00 70 [email protected] ◗ Provide a re-use low level solution Autosar and ISO2626-2 compliant. ◗ Allow industrial to increase their productivity when they change of hardware platform by providing a smart configuration. ◗ Work on RTE middleware (Run-Time Environment), in order to warrant a better flexibility for the application integration on low level platforms, and better performances. ◗ Integrate PharOS technology on Autosar Architecture. PARTNERS Large companies: AUTOLIV, DELPHI, PSA SMEs: GEENSYS, KEREVAL Research institutes, universities: CEA, ESEO, ESTACA, IRCCYN PROJECT DATA Coordinator: GEENSYS Co-label: ID4CAR Call: FUI STATUS - MAIN PROJECT OUTCOMES Start date: April 2010 The project is organized around the following Work-packages: Duration: 36 months ◗ Methodology: Implementation of Safety application (WP1), Functional Validation method (WP7). ◗ Tools: Low level Configuration (WP3), RTE generator (WP5). ◗ Technology: Low level modules development (WP2), FlexRay implementation (WP6), OSEK test /ISO2626-2 (WP4). Global budget (M2): 5.2 Funding (M2): 2.1 ◗ Demonstrator using PharOS technology. Automotive & Transports WG 33 01 AUTO:Mise en page 1 12/05/10 17:28 Page 34 Algorithms & data fusion for localisation and vision ODIAAC Odométrie Intégrée pour des Applications d'Aide à la Conduite en milieu urbain Vehicle geo-location is essential to several automotive safety applications: ◗ Assistance summoning and accident location warning; ◗ Driver notification of visible or non-visible hazards (hidden vehicles at crossroads, sensitive area approach, speed limitation). Geo-location sensors capabilities such as GPS present some drawbacks in urban area which cause major variation to location accuracy or even losses of continuity of service. ODIAAC project has proven that an integrated camera can be used as a complement to GPS based geo-location systems. COMPLETED PROJECT CONTACT François GASPARD CEA LIST +33 (0)1 69 08 46 84 [email protected] PARTNERS PROGRESS BEYOND THE STATE OF THE ART Nowadays vehicle geolocation systems (GPS receiver mainly) have shown their ability to meet the geolocation need only due to driver control on the vehicle. In case of Advanced Driving Assistance Systems, as specified by automobile industry, new criteria have to be met: ◗ cost / accuracy compromise: next generation satellite-based solutions accuracy should be of a meter. This accuracy is only theorical and hide issues of signal occultation (in tunnels) or multipaths (reflection of signals on buildings). Use of differential correction can improve accuracy but has the drawback to need an expensive network of reference stations (user cost on priced subscription); SMEs: DOTMOBIL Research institutes, universities: CEA LIST, LASMEA PROJECT DATA Coordinator: CEA LIST ◗ measure integrity: for a fully satisfying safety Call: ANR device for geolocation, the positioning system has to provide quality of service which results to a confidence indicator on measure. This confidence is not fully assured by GPS systems; Start date: May 2007 ◗ continuity of service: geolocation has to be pro- Duration: 30 months vided anytime in which satellite technologies are ineffective due to occultations. It is then necessary to propose an alternative to keep on localizing the vehicle during GPS losses. Global budget (M2): 1.7 ODIAAC is a hybrid geolocation prototype of lowcost on-vehicle sensors (GPS, odometer, camera), which provides on-demand accurate position together with.integrity indicator. MAJOR PROJECT OUTCOMES ◗ Patents: 1 ◗ Publications: 5 publications including 3 international. ◗ Experimentations: Realisation of an embedded demonstrator allowing quantitative evaluation of system thanks to on the ground reality. Automotive & Transports WG 34 Large companies: RENAULT Funding (M2): 0.7 01 AUTO:Mise en page 1 12/05/10 17:28 Page 35 Software tools and Methods Iterative Design Process for Self-Describing Real-Time Embedded Software Components Development of a design method for self-adaptive systems that are built out of components. In order to integrate the latter into a (possibly already running) system, the components need to describe their needs in terms of functional and non-functional requirements (self-descriptive components). Whereas the approach is domain independent, the main application area is the automotive domain (key domain of partner Fraunhofer ESK). ON GOING PROJECT CONTACT Ansgar RADERMACHER CEA LIST, laboratoire LISE +33 (0)1 69 08 38 12 [email protected] ◗ During design, we need to evaluate resource consumption of system parts and optimize the estimation of overall system resource requirements. ◗ During runtime, want to adapt resource allocation depending of estimated ressource consumption and safety criteria, discover environmental changes or discover/use services that appear dynamically. TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS The complexity of upcoming embedded systems can not be managed by traditional methods, as studied in the automotive domain, as exemplary application area. Selfadaptive systems and self-organization paradigms are promising approaches to tackle arising challenges. But no suitable design methodology for self-adaptive systems allowing design and evaluation of self-adaptive behavior exist: today’s systems are developed statically at design time and lack the ability to describe their structure and status at runtime. This is inevitable information for self-adaptive systems enabling these to identify the context and adapt adequately. This system adaptability needs to be well defined, especially in real-time embedded systems which impose timing and resource constraints on the system behavior which must in particular be meet during the adaptation process. STATUS - MAIN PROJECT OUTCOMES PARTNERS Research institutes, universities: CEA, FRAUNHOFER ESK, MUNICH PROJECT DATA Coordinator: FRAUNHOFER - CEA Call: ANR Start date: September 2009 Duration: 36 months Global budget (M2): 0.6 (CEA) 1.2 (CEA+Fraunh.) Funding (M2): 0.3 (CEA) 0.8 (CEA+Fraunh.) The project has been presented at the PICF (Programme Inter Carnot Fraunhofer) event end of November. Major use cases have been defined (first deliverable). Automotive & Transports WG 35 01 AUTO:Mise en page 1 12/05/10 17:28 Page 36 Dependability of monitoring systems Safety Check of Automotive Software & Hardware Architectures Application of ISO 26262 standard on an example case of motor control in an automotive application. ◗ Replacement of an architecture with external redundancy (2 microcontroller) by an architecture with internal redundancy (dual core microcontroller). ◗ Development of an analysis tool, of an HMI, of numerical simulation of the system and its components including fault injection. ON GOING PROJECT CONTACT Jochen LANGHEIM STMICROELECTRONICS +33 (0)1 58 07 75 25 [email protected] ◗ Adaptation of component models, in particular of the dual core microcontroller. ◗ Creation of booklet with rules for a simple and rapid application of the ISO standard whilst assuring functional safety of the system. PARTNERS TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS ◗ Harmonise the approaches of functional safety in the supply chain considering: • ISO 26262 (especially the creation of the safety files); • AUTOSAR; • System design. Large companies: DELPHI, RENAULT, STMICROELECTRONICS Intermediate size enterprise: ESG FRANCE SMEs: KNOWLEDGE INSIDE ◗ Thanks to the innovative tool of Know- Research institutes, universities: UTC ledge Inside it shall be possible to: • Better understand the market needs; • Develop consequently an offer; • Become more competitive. ◗ Improve the understanding between PROJECT DATA the “system” actors and the “component” actors. Coordinator: STMICROELECTRONICS STATUS - MAIN PROJECT OUTCOMES Co-label: MOVEO ◗ A tool environment allowing the description of heterogenous systems and the inte- Call: FUI8 grated safety analysis in systems developments (Version DEMO architect). ◗ Models harmonised between actors. ◗ A well explained application case for the industry. ◗ A complete solution for the introduction of the ISO26262 at the officialisation of the standard in 2011. ◗ Drafts of templates facilitating the collaboration between the actors in the supply chain (Datasheet). Start date: October 2009 Duration: 24 months Global budget (M2): 4.4 Funding (M2): 1.4 Related Systematic project(s): EDONA Automotive & Transports WG 36 01 AUTO:Mise en page 1 12/05/10 17:28 Page 37 Electronics Architecture Systèmes Critiques pour l’Automobile : Robustesse des Logiciels Exécutifs Temps-réel The main objective of the SCARLET project is to build methods and software techniques for the optimized implementation of robustness mechanisms in multi-layered automotive software supporting future highly critical functions. The project will take into account automotive standards and practice such as AUTOSAR, future ISO26262 and the use of Off The Shelf Components, commercial or not. The project consists in three main work packages. ◗ The first work package defines a methodology indicating the requirements, criteria and recommendations for the use of software robustness mechanisms. ◗ The second work package aims at developing mechanisms for software robustness improvement for open or closed (COTS) software. ◗ And the third one consists in characterizing and validating the methodology and mechanisms on a demonstrator derived from a vehicle. PROGRESS BEYOND THE STATE OF THE ART ◗ Requirements and industrial needs, constraints of automotive domain. • State of the art by the academics. • Fault model developed and shared. ◗ Definition of the reflexive approach and experimentation going on an experimental Autosar platform. ◗ Runtime environment for hard real-time constraints for automotive. First deliverable and mockup. ◗ Taking into account the automotive context: • Autosar (study of mechanisms of latest specifications); • Under construction ISO 26262; • Interaction with silicon vendors: Freescale, Infineon. ◗ Runtime environment real-time and robust for automotive [CEA LIST]. • Dependability - Generic (application-independent mechanisms); - Reproducibility of the execution; - Automatic partitioning spatial and temporal; - Management of failure allowing the recovery of only the defective function without interference on the rest of the system. • Coexistence of different levels of criticality possible. • Requirements of industrials and ISO 26262 (ASIL D) addressed => Mockup made on chip S12XEP100. COMPLETED PROJECT CONTACT Philippe QUERE RENAULT +33 (0)1 76 85 63 60 [email protected] PARTNERS Large companies: RENAULT, VALEO SMEs: TRIALOG Research institutes, universities: CEA, CNRS-LAAS, INRIA-LORIA, IRCCYN PROJECT DATA Coordinator: RENAULT Call: ANR2006 Start date: May 2007 Duration: 42 months MAJOR PROJECT OUTCOMES ◗ Publications: International conference • Robustness of modular multi-layered software in the automotive domain: a wrapping-based approach. Espagne, 22-26 Septembre 2009. • An approach for improving fault-tolerance in automotive modular embedded software. 17th International Conference on Real-Time Networked Systems (RTNS 2009), Paris, 26-27 Octobre 2009. • How to configure AUTOSAR OS timing protection. Paris, France, October 2009. • An analysis of the AUTOSAR OS timing protection mechanism. Spain, September 2009. • Temporal isolation for the cohabitation of applications in automotive embedded software. Valencia, Spain, April 2010. • A Spatial and Temporal Partitioning Approach for Dependable Automotive Systems. Majorque (Espagne), Septembre 2009. • Towards optimal priority assignment for probabilistic CAN-based systems. WIP session of WFCS'2010, March 2010. • Towards optimal priority assignment for probabilistic real-time systems with variable execution times. Paris, October 2009. • Preliminary results for introducing dependent random variables in stochastic feasiblity analysis on CAN. Dresden, May 2008. • Probabilistic real-time schedulability: from uniprocessor to multiprocessor when the execution times are uncertain. Thèses de doctorat • Robustesse du logiciel embarqué multicouche par une approche réflexive: application à l'automobile. Global budget (M2): 3.7 Funding (M2): 1.5 Automotive & Transports WG 37 01 AUTO:Mise en page 1 12/05/10 17:28 Page 38 Diagnosis SEEDS Smart Embedded Electronic Diagnosis Systems Automotive electronic system's complexity is growing. With up to 4 km of wires, the importance of the electrical harness in modern cars is improving. New functions (by wire systems, in wheel motors, fuel cell) in the near future will make the harness a critical function on its own. The objective of the SEEDS project is to design and develop an automatic diagnosis tool for car harnesses, both for production or garage maintenance and for in-use diagnosis; able to detect - localize and characterize defects in a complex topology harness. COMPLETED PROJECT CONTACT Marc OLIVAS CEA-LIST +33 (0)1 69 08 48 83 [email protected] PROJECT RESULTS PARTNERS ◗ Products: Large companies: DELPHI, RENAULT TRUCKS, SERMA INGENIERIE Cable or harness diagnosis system for external maintenance application (garage) applied to cars and trucks. SMEs: MONDITECH ◗ Patents: • FR 06 06531 – Procédé et dispositif d’analyse de réseaux de câbles électriques (WO/2008/009566). • FR 08 50013 – Procédé pour l’amélioration de la précision de détection et de localisation de défauts par réflectométrie dans un réseau électrique câblé. Research institutes, universities: CEA, LGEP, INRIA ◗ Technologies: Cable and harness diagnosis method based on reflectometry, application to car and trucks harnesses. • Propagation models of diagnosis signals in cables. • Diagnosis theory and models: inverse scattering for defect detection and localisation, model inversion using genetic algorithm. ◗ Publications: ◗ Experimentations: The external diagnosis prototype has been packaged and sent to Delphi's location in Warwick (England) where it has been under test in real garage conditions since october 2008. During this test phase, several cars and trucks have been prepared with various cases of harness defects. Both the project's representative and the mechanics from Delphi have used the diagnosis system to find the defects and repair them efficiently. Befor the system was available, usual time needed to repair such harness defects could reach 2 days, now the average time is less than an hour. SEEDS using the system Automotive & Transports WG 38 Coordinator: CEA LIST Start date: February 2006 Duration: 36 months 10 publications. SEEDS system under test PROJECT DATA Global budget (M2): 3.1 Funding (M2): 1.6 01 AUTO:Mise en page 1 12/05/10 17:28 Page 39 Electronics Architecture / Dependability of monitoring systems SIRSEC Système d’Information Reparti Sécuritaire Driverless railways transport systems exhibit the growing need for information exchange which increases with every new generation of systems. The need of information exchange has to be conciliated with the severe safety constraints imposed on these systems. Under the pressure of these constraints, system developers use almost uniquely proprietary ad hoc solutions which are reputated to be more easily certifiable against stakeholders' requirements than the solutions based on open standards. It is proved that the cost of acquisition and maintenance of open system is in general lower than the cost of proprietary solutions. SIRSEC project proposes through the development of Design Patterns a methodological and technological basis which allows to use open solutions but preserve the possibility of evaluation of safety with relative ease. TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS SIRSEC project objective is to develop process, technologies and tools for the safety of information exhange. The safety level of these exchanges and services must be compatible with the SIL (Safety Integrated Level) of applications using these informations. ◗ The project aims to define an architectural platform for data exchange for shared applications defined from SIL0 to SIL4. ◗ The major constraint for the envisaged applications consists in the maintain of safety characteristics without sacrifying operational performance and diversity of implementation environments. The proposed solution is based on the safety featured middleware platform. ON GOING PROJECT CONTACT Lionel SEGUY ALSTOM TRANSPORT +33 (0)4 72 81 52 75 lionel.seguy @transport.alstom.com PARTNERS Large companies: ALSTOM TRANSPORT, SERMA INGENIERIE, THALES SMEs: GEENSOFT Research institutes, universities: CEA, INRETS, IRIT ◗ The industrial context of the project takes into account several constraints unachieved by todays technologies used in the railway applications. These constraints put into conflict two main factors which determine application possession costs: the cost of production and the cost of commissionning and maintenance of the solution. In clear terms, systems with high level of dependability which should be proved and certified use almost exclusively proprietary solutions strongly oriented towards application. The major commercial objective for the consortium members is the access to the market of solutions of high level of dependability. This will be achieved in the case of evolution of currently proposed products towards the incorporation of open standards for distributed safety critical applications. STATUS - MAIN PROJECT OUTCOMES The major constraint for the envisaged applications consists in the maintain of safety characteristics without sacrifying operational performance and diversity of implementation environments. The proposed solution is based on the safety featured design patterns of different middleware platforms. PROJECT DATA Coordinator: ALSTOM TRANSPORT Call: FUI7 Start date: September 2009 Duration: 36 months Global budget (M2): 4.6 Funding (M2): 1.9 Automotive & Transports WG 39 01 AUTO:Mise en page 1 12/05/10 17:28 Page 40 Algorithms & data fusion for localisation and vision SPEEDCAM Speed limit determination using camera and maps The aim of the project is the development of a multisensory fusion system combining embedded frontal vision system and a Geographic Information System for the determination of the current speed limit. This so-called "Pre-set ACC" will be built on real demonstrators (prototype vehicles) belonging to industrial partners. TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS ON GOING PROJECT CONTACT Fawzi NASHASHIBI ARMINES +33 (0)1 40 51 92 31 Fawzi.Nashashibi@ mines-paristech.fr ◗ Design and development of a single embedded camera vision system ◗ Detection/recognition in all circumstances of all speed limit sign boards and supplementary signs ◗ Development of a new multisensor data fusion algorithm combining vision and GIS providing the speed limit and its associated confidence index ◗ Production of "Bench tests" and number of relevant databases PARTNERS Large companies: DAIMLER, VALEO Research institutes, universities: ARMINES - HOCHSCHULE AALEN ◗ Integration of the functionality in an equipped vehicle ◗ Realization of a new ADAS system STATUS - MAIN PROJECT OUTCOMES PROJECT DATA Coordinator: ARMINES ◗ Scientific and technical repercussions: • Development of new real time techniques for traffic signs recognition in all conditions; • Same approaches could be exploited for other traffic beacons; • Generic fusion approach: could be used in other driving assistance systems. Co-label: MOV'EO Call: ANR ◗ Societal and environmental impacts: • Accident reduction: new driving aids; • Less traffic congestion; • Pollution reduction; • Enhance transport efficiency; • Stress reduction; • Less speeding tickets due to inattent. Start date: October 2009 Duration: 36 months Global budget (M2): 3.6 Funding (M2): 0.7 Automotive & Transports WG 40 01 AUTO:Mise en page 1 12/05/10 17:28 Page 41 Dependability of monitoring systems SYSPEO SYStem Proof Extended Objective ON GOING PROJECT Syspeo objective is to extend the usage of formal methods for the development of mechatronic systems in the automotive industry. The ultimate goal is to enhance products quality with reduced iteration in the development and validation cycle. The challenge is to deploy formal proof based on system requirements on hybrid models mixing discrete and continuous domains. ◗ Formal monitoring of complete mechatronic models executed as black boxes. ◗ Capability to use formal proof of system requirements on complete mechatronic sys- CONTACT Arnaud BALMITGERE BWI France +33 (0)1 49 90 42 56 [email protected] tems models. ◗ Generation of scenarios for embedded software validation corresponding to discrete model. PARTNERS TECHNOLOGICAL OR SCIENTIFIC INNOVATIONS Large companies: CEA-LIST, UCL, WOW ◗ Modelling of a controlled chassis system of a vehicle, using symbolic description techniques of complex multibody systems, leading to a differential equations system representation. ◗ Specification and implementation of algorithms to monitor simulations of models with respect to a formalized set of system safety properties. ◗ Hybrid (discrete & continuous) model parsing, leading to an approximated set of reachable system states. ◗ Formal proof on parsed hybrid model. ◗ Enrich system validation with automatic generation of scenario. ◗ Formal proof methodology benchmarking, versus real system test bench validation. PROJECT DATA Coordinator: BWI FRANCE Call: EUREKA Start date: September 2007 Duration: 30 + 12 months STATUS - MAIN PROJECT OUTCOMES Global budget (M2): 2.4 ◗ Project started in September 2007. ◗ System modelling complete, further refinements are being introduced. ◗ Monitoring algorithms specification and implementation complete, second genera- Funding (M2): 1.2 tion of the tool is in development. ◗ System properties definition and implementation of the verification experiment in process. ◗ Hybrid model parsing complete. ◗ System test bench definition and implementation complete. Automotive & Transports WG 41