ing to BMW, electronic components comprise more than 30 percent of a car’s manufacturing cost. The trend in the car manufacturing industry is to acquire more in-house elec- tronics competence to capture added value that previously went to subsystem suppliers. The strategy calls for software and hardware standards that will facilitate plug-and-play subsystems, reducing the strategic importance of any single subsystem supplier. The Offene Systeme und deren Schnittstellen für die Elek- tronik im Kraftfahrzeug (open systems and corresponding interfaces for automotive elec- tronics), or OSEK, operating system require- ments are an example of this policy.2 Clearly, however, without an overall understanding of the interplay of subsystems and the difficul- ties of integrating highly complex parts, sys- tem integration is increasingly a nightmare for car manufacturers. In addition, subsystem suppliers are trying to enlarge their perimeter of competence to capture more added value.
Automobile electronics comprises three basic domains:
power train management—for example, electronic control units (ECUs) that con- trol ignition timing and the amount of fuel injected into the cylinders;
body electronics—for example, ECUs that control dashboard displays, suspen- sion settings, and temperature; and
information processing, communication with the outside world, and entertain- ment (often called the telematics or info- tainment system).
The first domain is typical of any trans- portation system and is characterized by tight safety and efficiency constraints. Its core com- petence is control algorithms, along with soft- ware and mechanical-electrical hardware design and implementation.
The body control domain involves the management of a distributed system that increasingly resembles a network with proto- cols likely to have different requirements than standard communication protocols. Guaran- teed services are the essence of this domain.
A car’s infotainment system is the product of industrial domains that are progressing in the technology race at a faster rate than the automotive domain. This domain can reap
the most short- to medium-term profits. Cus- tomers now often base buying decisions on the infotainment environment more than engine performance and handling. Hence, the question arises: What will constitute a car company’s core competence? Will the elec- tronic components be the car and the mechanical components an accessory?
For car manufacturers, system design is def- initely the most important technology they must master to improve the quality and increase the value of their cars’ electronic com- ponents. Designers of automotive electronic systems need a methodology that focuses on two main principles: separation of concerns and platform-based design.
Electronic-system design issues
To support the electronic-design chain, sys- tem designers in the automobile industry must establish a new design flow. Clean inter- faces and unambiguous specifications are essential parts of this design flow. In addition, the design flow must address the thorny issue of IP protection. This is even more important in the automotive domain than in other industrial segments because the automotive- supplier chain is deeper.
The following issues are likely to determine the preferred approaches to the design and implementation of complex embedded sys- tems in the automotive domain (and others):
Reuse. Design time and cost will dominate sys- tem designers’ decision-making process. Therefore, design reuse of all kinds, as well as just-in-time, low-cost design debugging tech- niques, will be highly important. Design flex- ibility is essential to mapping an ever-growing functionality onto a continuously evolving set of associated hardware implementation options.
High levels of abstraction. Designers must cap- ture designs at the highest abstraction level to exploit all the available degrees of freedom. This abstraction level should not distinguish hardware from software, since this distinction is the consequence of a design decision.
Concurrency. The implementation of efficient, reliable, and robust approaches to the design, implementation, and programming of con-