Search Results

Model-based software development and automatic production code generation have become increasingly established in recent years. The aerospace industry and other industries, such as automotive, have widely adopted and successfully deployed these methods in many different series production programs worldwide. This brought various benefits, such as a reduction in development times and improved quality due to more precise specifications, and early verification and validation by means of simulation. Model-based development is a general purpose development approach which can be applied to a wide variety of applications. Safety-critical systems, like found in aerospace applications to a large extent, but also found increasingly more often in other industries, like automotive or medical devices, pose special additional requirements to this process. This paper describes how model-based design and automatic production code generation can be applied to the development of safety-critical software.

Model-based development and autocoding have become common practice in the automotive industry over the past few years. The industry is using these methods to tackle a situation in which complexity is constantly growing and development times are constantly decreasing, while the safety requirements for the software stay the same or even increase. The debate is no longer whether these methods are useful, but rather on the conditions for achieving optimum results with them. From the experiences made during the last decade this paper shows some of the key factors helping to achieve success when introducing or extending the deployment of automatic code generation in a model-based design process.

This paper will first introduce and classify the basic principles of architecture-driven software development and will briefly sketch the presumed development process. This background information is then used to explain extensions which enable current behavior modeling and code generation tools to operate as software component generators. The generation of AUTOSAR software components using dSPACE's production code generator TargetLink is described as an example.

The design of a complex real-time embedded system requires the specification of its functionality, the design of the hardware and software architectures, the implementation of hardware and software components and finally the system validation. The designer, starting from the specification, refines the solution trying to minimize the system cost while satisfying functional and non functional requirements. The automatic code generation from models and the introduction of the platform-based design methodology can drastically improve the design efficiency of the software partition, while maintaining acceptable the cost overhead of the final system. In this approach, both top-down and bottom-up aspects are considered and solutions are found by a meet-in-the-middle approach that couples model refinement and platform modeling.

In future cars, mechanical and hydraulic components will be replaced by new electronic systems (x-by-wire). A failure of such a system constitutes a safety hazard for the passengers as well as for the environment of the car. Thus electronics and in particular software are taking over more responsibility and safety-critical tasks. To minimize the risk of failure in such systems safety standards are applied for their development. The safety standard IEC 61508 has been established for automotive electronic systems. At the same time, automatic code generation is increasingly being used for automotive software development. This is to cope with today's increasing requirements concerning cost reduction and time needed for ECU development combined with growing complexity. However, automatic code generation is hardly ever used today for the development of safety-critical systems.

Permanently increasing software complexity of today's electronic control units (ECUs) makes testing a central and significant task within embedded software development. While new software functions are still being developed or optimized, other functions already undergo certain tests, mostly on module level but also on system and integration level. Testing must be done as early as possible within the automotive development process. Typically ECU software developers test new function modules by stimulating the code with test data and capturing the modules' output behavior to compare it with reference data. This paper presents a new and systematic way of testing embedded software for automotive electronics, called MTest. MTest combines the classical module test with model-based development. The central element of MTest is the classification-tree method, which has originally been developed by the DaimlerChrysler research department.