Ines Fey

Systematic Model-Based Testing of Embedded Automotive Software

The software embedded in automotive control systems increasingly determines the functionality and properties of present-day motor vehicles. The development and test process of the systems and the embedded software becomes the limiting factor. While these challenges, on the development side, are met by employing model-based specification, design, and implementation techniques, satisfactory solutions on the testing side are slow in arriving. With regard to the systematic test design and the description of test scenarios especially, there is a lot of room for improvement. This paper introduces the model-based black-box testing (MB^3T) approach in order to effectively minimize these de cits by creating a systematic procedure for the design of test scenarios for embedded automotive software and its integration in the model-based development process. According to the MB^3T approach, logical test scenarios are rst de ned based on the textual requirements specification of the embedded software. These test scenarios are specified at a high level of abstraction and do not contain any implementation details of the test object. Due to their close link to the requirements it is easy to check which requirements are covered by which test scenario. Subsequently, the requirement-based logical tests are re ned to executable model-based test scenarios. Finally, the approach helps to check, whether or not the logical test scenarios are fully covered by the executable test scenarios. The MB^3T approach has recently been successfully employed in a number of automotive embedded software development projects at DaimlerChrysler.

Experiences with model and autocode reviews in model-based software development

Through the introduction of the model-based development, paradigm models became first class citizens in the development of in-vehicle software and are thus also object to strict quality assurance. Just as code reviews are widespread in classical software development, models also have to undergo a stringent review procedure - particularly if they serve as a basis for automatic software implementation by means of model-based code generators. In addition to model reviews, the generated production code (autocode) must be reviewed by performing so-called autocode reviews. This paper presents our procedures for a combined model and autocode reviews and provides examples, lessons learned, as well as significant experimental results drawn from a typical automotive embedded software development project.

Modellbasierte Entwicklung eingebetteter Fahrzeugsoftware bei DaimlerChrysler

Steigende Softwareumfänge und die damit einhergehende Komplexitätssteigerung erfordern neue Herangehensweisen an die Entwicklung eingebetteter Fahrzeugsoftware. Als Reaktion auf diese Herausforderungen vollzieht sich seit einigen Jahren ein Paradigmenwechsel von der klassischen Programmentwicklung hin zu modellbasierten Techniken. Der Beitrag gibt einen Überblick über den Einsatz modellbasierter Entwicklungstechniken bei DaimlerChrysler und zeigt aktuelle Forschungsschwerpunkte auf. AbstractThe growing amount of software and attendant rise in complexity demand new approaches to the development of embedded vehicle software. For some years now, a paradigm change from classical program development to model-based techniques has been taking place in reaction to these challenges. This paper provides an overview of the application of model-based development techniques at DaimlerChrysler and presents the main foci of current research.

Automatic Test Generation for Model-Based Code Generators

This paper presents a novel testing approach for model-based design tools, termed GraphOTK, and applies it to the optimizing component of a code generator for Simulink and Stateflow models.

Quality Assurance Methods for Model-Based Development: A Survey and Assessment

This paper examines state-of-the-art quality assurance (QA) techniques for model-based software development in the automotive domain. Both the aims and effort required to apply a certain method are discussed for all relevant QA techniques. Since QA techniques can only be used effectively if they are seamlessly integrated within the overall development process and among each other, an appropriate interconnection and order of application is important. Based on our experience from automotive software development projects, we suggest a QA strategy that includes the selection of QA techniques and the sequence of their application.

TestML - A Test Exchange Language for Model-Based Testing of Embedded Software

Test processes in the automotive industry are tool-intensive and affected by technologically heterogeneous test infrastructures. In the industrial practice a product has to pass tests at several levels of abstraction such as Model-in-the-Loop (MIL), Software-in-the-Loop (SIL) and Hardware-in-the-Loop (HIL) tests. Different test systems are applied for this purpose (e.g. dSPACE MTest, dSPACE Automation Desk, National Instruments Teststand) and almost each test system requests its own proprietary test description language. The exchange of tests between different test systems and the reuse of tests between different test levels is normally not possible. Efforts to integrate these heterogeneous test environments, to address test exchange in a general manner and to standardize and harmonize the existing language environment are still at the beginning and not tailored towards the requirements of the automotive domain. To keep the whole development and test process efficient and manageable, the definition of an integrated and seamless approach is required. TestML --- the test exchange language we present in this article --- is defined to overcome the technological obstacles (different test language syntax and semantics, different data formats and interface descriptions) that almost automatically accompany the application of heterogeneous test tools and test infrastructures. TestML supports the exchange of tests between different test notations in a heterogeneous tool environment. In this paper, we introduce the XML schema of TestML and demonstrate the efficiency of the interchange format by giving examples from the model-based development of electronic control units. Tool support is illustrated by an application with Simulink/Stateflow.

Systematic Unit-Testing of Ada Programs

The systematic test is an inevitable part of the verification and validation process for software. Overall support for all testing activities is currently not available in a single Ada testing tool. Hence, a combination of powerful testing tools is necessary to provide systematic and complete test process automation for the examination of Ada programs. The classification-tree editor CTE supports the systematic design of functional test cases. The strengths of AdaTEST are the comprehensive support for test execution and coverage analysis. The combination of both tools leads to systematic and well-documented test procedures. It has already been successfully applied to several real world examples including aerospace applications.

Using Model and Code Reviews in Model-based Development of ECU Software

In the automotive industry, the model-based approach is increasingly establishing itself as a standard paradigm for developing control unit software. Just as code reviews are widespread in classical software development as a quality assurance measure, models also have to undergo a stringent review procedure – particularly if they serve as a starting point for automatic implementation by code generators. In addition to these model reviews, the generated production code is reviewed later in the development process by performing auto code reviews. This article will present procedures for and give an account of experiences with model and code reviews which have been adapted to the model-based development process.