Computeraided Architecture Design and Optimized Implementation of Distributed Automotive EE Systems - PowerPoint PPT Presentation


PPT – Computeraided Architecture Design and Optimized Implementation of Distributed Automotive EE Systems PowerPoint presentation | free to download - id: 7cce0-MzA1N


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation

Computeraided Architecture Design and Optimized Implementation of Distributed Automotive EE Systems


Basic Software (BSW) implementation of the AUTOSAR infrastructure (nearly 40 ... Basic AUTOSAR concepts (Cont. ... Instantiation of basic elements of the ... – PowerPoint PPT presentation

Number of Views:218
Avg rating:3.0/5.0
Slides: 29
Provided by: antalr
Learn more at:


Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Computeraided Architecture Design and Optimized Implementation of Distributed Automotive EE Systems

Computer-aided Architecture Design and Optimized
Implementation of Distributed Automotive EE
  • Ajay Kumar, Antal Rajnak
  • Automotive ProgramsSystem Level Engineering

June 06, 2007 DAC 07, Paper 31.4
Presentation Outline
  • Problem statement
  • Basic AUTOSAR concepts
  • Proposed Tool Flow
  • Experience thus far
  • Suggested improvements / future work
  • Summary

Current Design Environment
  • Manual process with inadequate tool support
  • Insufficient data to support early decision
    making, leading to early binding and late
  • ECU-focused rather than system-level
  • Validation on physical prototypes
  • Distributed, multi-company, modular supply chain
  • Few if any widely-accepted standards
  • The resulting architectures are seldom optimal in
    terms of
  • Performance
  • Flexibility / Scalability
  • Reliability
  • Life-cycle Cost

Industry trends
  • Increasing use of model-based function
  • 20-25 of all EE functions covered today
  • Used for
  • Functional verification
  • Executable specification OEM -gt Tier1
  • Major standardization effort - AUTOSAR
  • Version 2.1 of specs released matured
  • Use gaining momentum in Europe
  • US and Japan watching closely
  • Step towards development and validation in a
    virtual environment, using Systems Engineering

Basic AUTOSAR concepts
  • AUTomotive Open System ARchitecture
  • Layered software architecture, with clearly
    defined interfaces
  • Applications implemented as atomic or composed
    Software Components (SWC)
  • Virtual Function Bus (VFB) abstraction of all
    interconnected SWCs, communicating exclusively
    through ports independent of underlying hardware
  • Run Time Environment (RTE) implementation of
    the VFB on a particular ECU
  • Basic Software (BSW) implementation of the
    AUTOSAR infrastructure (nearly 40 different

AUTOSAR ECU sw structure source
Basic AUTOSAR concepts (Cont.)
  • AUTOSAR systems are described by a set of XML
    based Description Files
  • Derived from a defined UML metamodel
  • Whats described ECUs, SWCs, their mapping to
    ECUs, ECU interconnections etc.
  • Whats missing Methodology to create these
    description files

The Implementation Abyss
  • Model based function design a widely
    accepted,and used technique
  • AUTOSAR brings new level of abstraction to
    function implementers
  • But
  • The task to translate a model based design into
    a robust, and efficient system, in a highly
    distributed environment remains a challenge.
  • How to bridge the implementation abyss?

Proposed SLD methodology
  • Architecture Design
  • Topology definition
  • SWC to ECU Vehicle function creation
  • mapping
  • VFB-level simulation
  • Bus scheduling
  • Initial ECU Scheduling
  • Metrics generation
  • ECU configuration
  • Runnables to task mapping
  • OS task scheduling
  • Configuration of remaining BSW modules

Proposed System Level Tool Chain
  • Implemented as a set of well connected
  • Vehicle System Architect for the OEMs
  • Function Editor
  • Topology Editor
  • Function Mapper
  • VFB-level Simulator
  • Network Cluster Builder
  • ECU Scheduler
  • Harness Design Tool (Capital Harness)
  • Metrics Generator
  • Vehicle System Builder for the Tier 1s

Proposed Tool Flow
Tool Implementation Challenges
  • Support for concept evaluation for EE-systems in
    early phases of the design process, when the
    available level of detail is minimal
  • Connectivity to legacy OEM environments
  • Scalable tool functionality, depending on OEM
  • Manual / Interactive use mode
  • Analysis
  • Synthesis requirements based design

Vehicle Function Editor
  • Users create atomic or composed SWCs representing
    vehicle functions
  • Method
  • Manual graphical entry
  • Import from function modeling tools(e.g.
  • Envelope of function is adequate
  • Interface consistency is checked

Vehicle Topology Editor
  • Instantiation of basic elements of the physical
    architecture (ECU-s, sensors, actuators, buses
  • Allows creation of soft (only partially
    defined) components in the early phases of
  • Import
  • Individual components
  • Pre-existing message matrix
  • Pre-existing topology
  • Graphically arrange components to form desired
    physical architecture.

Topology editing Function mapping
Vehicle Function Mapper
  • Function
  • Mapping of vehicle functions to logical ECUs
  • Check of resource demand vs. availability
  • Soft ECUs allocated to real ECUs from library
  • Split of end to end timing requirements btw ECUs
    and buses
  • Automatic gateway configuration
  • Output
  • System Configuration Description file
  • ECU extract of SCD
  • Export to Harness design tool

VFB-level Simulation
  • VFB-level view network of SWCs connected through
  • This network can be simulated
  • Appropriately compiled SWCs with behavior
  • Verify desired functionality across single, as
    well as multiple ECUs

Network Scheduling the Cluster Builder
  • Basic Functionality
  • Produces schedule tables
  • Creation/modification of frames (PDU-s) and
    related parameters (ID, period time, x-fer mode
  • Interactive signal packing and editing of
    schedule tables
  • Cluster schedulability analysis
  • Expanded functionality
  • Automatic frame compilers (AFC-s) for
  • Fully automatic scheduling and frame packing

ECU Scheduler
  • Runnables to tasks mapping
  • Manual configuration and schedulability analysis
    based on available timing information
  • Configuration of OS and BSW schedulers
  • Tightly coupled to network scheduling.

Unified Timing Model
Metrics generation
  • Compares and ranks a set of alternative
    architectural solutions in relative terms
  • Component cost is only one aspect - scalability,
    flexibility and extensibility will heavily
    influence the results.
  • Scenario based sensitivity analysis is essential
    to identify decision points
  • Simplified analysis running in the background
    provides live feedback to user

Vehicle System Builder
  • Function
  • Create/edit all AUTOSAR-relatedconfiguration
  • Intelligently merge partial configuration
    filesfrom multiple sources
  • Generate a downloadable binary image of the
    complete software (AppsBSWConfig) for the
    selected ECU.

Vehicle System Builder
Correctness by Design
  • Vision
  • Requirements based Systems Engineering Process
  • Holistic, top-down approach, supported by well
    connected high-level tools.
  • Development and validation in a virtual
  • Values
  • Reduced cost
  • Shortened lead-times
  • Improved Reliability Quality

Experience thus far
  • Tool-set for requirements based synthesis of
    complex CAN and LIN networks and gateways
    successfully deployed and used in production
    since 1998 (Volcano Network Architect tool).
  • Users are enthusiastic!
  • But...
  • Timing analysis is not widely used in current
    automotive design flows.

Suggested improvements / future work
  • Enhanced AUTOSAR Timing Model to cover
    important concepts, i.e. jitter, phasing, or
    precedence required to deal with complex systems
  • Coordinated AUTOSAR Metamodel extensions
  • Evaluation metrics refinement
  • Define objective metrics and evaluationalgorithms
    to judge
  • Scalability
  • Flexibility

  • Bridging the implementation abyss between model
    based function design and creation of efficient
    AUTOSAR implementations calls for new methods,
    and advanced tool support.
  • Challenges
  • Usability in early project phases, when only
    limited amount of data is available
  • Need to support a flexible set of usage
    models(interactive, analytic or full synthesis)
  • The required technology is readily available
  • The AUTOSAR Timing model needs major re-work


  • Thank you!