With the rising complexity of electronics in vehicles, system-level EDA tools have begun to play a greater role in automotive designs. As a result, all major EDA tools suppliers have thrown their towels in the ring. Recognizing this need, this year's 44th Design Automation Conference (DAC) in San Diego, Calif., June 4-8, has a special theme, “Automotive Systems.” While DAC is more of a silicon design conference, automotive is more than silicon design as it involves several disciplines.
To better understand the challenges confronting the EDA industry, a series of events have been organized by Nick Smith, Mentor Graphics' product line director of Integrated Electrical Systems Division. The conference has invited Lawrence D. Burns, vice president of research & development and strategic planning for General Motors Corp., to deliver a keynote address. Burn's talk, “Designing a New Automotive DNA,” will emphasize that future vehicles will need a new DNA that exchanges the internal combustion engine, petroleum, and mechanical linkages for fuel cells and batteries, hydrogen and electricity, and electronic systems and controls. This paradigm shift was first suggested at last year's Convergence (see "Reinventing the Automobile with New DNA," Auto Electronics, November 2006.
June 6 is devoted to technical sessions, panel discussions and lectures. The program starts at 8:30 a.m. with a discussion of automotive network design and analysis. Four papers shed light on network protocols such as CAN and FlexRay and investigate issues that remain to be resolved to model safe yet efficient distributed architectures. The first three papers present work on period synthesis, timing analysis and jitter relevant to drive-by-wire applications. The fourth paper discusses a model-checking approach to schedule optimization.
Next, a panel discussion will explore fully autonomous passenger vehicles that conduct navigation missions in urban environments. The goal of the Urban Challenge is to develop vehicles that can safely drive themselves in realistic urban settings. To succeed, the vehicles must obey traffic laws while safely merging into moving traffic, driving through traffic circles and busy intersections, and parking. A multimedia presentation and demonstration of Intel/Carnegie Mellon University's entry into the Urban Challenge will provide an example of current results.
Following the panel will be an invited session, moderated by David W. Smith of Synopsys. The first presentation will highlight the differences in automotive silicon design practices vs. other application segments, including commentary on the EDA requirements. The second paper will focus on creating robust designs by modeling safe operating areas in two hardware description languages. The third describes leading-edge research into autonomous vehicles: cars that can safely drive themselves in realistic urban environments.
Embedded software is a big part of automotive design. Despite such advances, there is a gap between how parts suppliers (hardware, software) to the automotive industry test their products vs. what a customer expects. Not only do the end products need to be validated, but so do the tool chains used to create the products. Left to the current state of the art, this is going to be a limiter to the growth in the industry. The goal of this panel is to explore the need for correctness and reliability in automotive software.
A special session will feature experts in the domain of virtualization mechanisms used to reduce the number of hardware processors in automotive electronics and users of the technology. Traditionally, adding new information-processing functionality in cars required the addition of computing hardware, leading to unacceptable growth in the number of microcontrollers. Virtualization of resources will be used to implement different applications on shared physical hardware. Applications will run on virtual platforms, appropriately shielded from other applications through adequate protection mechanisms.
Designing for a virtual automotive platform involves the interplay of several factors and impacts all aspects of safety-critical distributed embedded system design including hardware, software, applications, real-time scheduling and communication protocols.
Finally, panelists will debate the impact of the one-ton mobile platform on EDA tools.