What Would Joe Do?
Peggy Aycinena is a freelance journalist and Editor of EDA Confidential at www.aycinena.com. She can be reached at peggy at aycinena dot com.
Bosch Kibosh: Ask not what EDA can do for tomorrow’s cars
November 19th, 2014 by Peggy Aycinena
Sometimes shocking news arrives quietly, like fog over the Bay. In the very first sentence of his glamorous November 3rd keynote at ICCAD in Silicon Valley, Bosch VP of Engineering Peter van Staa bluntly told his audience that EDA would “probably not” help solve the problems related to electronic designs for the cars of the future. He then spent the next 45 minutes explaining instead how Bosch would.
“Bosch is unique,” van Staa said. With 42,000 employees and 4.5 billion euros spent in 2013 alone for R&D across multiple silos – energy and building technologies, consumer goods, industrial controls, and mobile “solutions” [anything driven by an engine being a particular focus] – the Stuttgart-based company is spewing out “an invention every 25 minutes.” [Can any EDA company make such a claim?]
The keynote continued with a detailed discussion of Bosch’s significant market position in MEMS for automotive, and how such power will propel the company into a leadership role building the cars of the future, specifically the self-driving ones.
“The car is the most complex system [under exploration] today,” van Staa said. With 150+ micro-controllers and so many ASICs on board, 80-percent of semiconductor innovation today is related to the car, he added, even though only 10 percent of production worldwide is targeted at automotive.
And no matter if the power train involves internal combustion or electric batteries, van Staa said the problems of automated driving are the same: They center around navigating from Point A to Point B, latency issues in communication between vehicle and critical Cloud-based decision-making, and vehicle reaction time to sudden obstacles in the road, other drivers who are incompetent and small children who dash out into the street without warning.
At this point, the audience was treated to a Bosch-produced video extolling the glamor and glory of the self-driving autos the company’s developing. Enigmatically filmed here in the Bay Area, in Alameda on the now-abandoned Naval Air Field, the film included young starry-eyed Bosch engineers watching with unmitigated joy as their self-driving cars were put through their paces on the tarmac.
The film also included footage of other young Boschers surfing the web on mobile devices or reading newspapers while seated confidently behind the wheel of an auto on auto-pilot, with fast images scrolling past outside the windows to indicate that said vehicle was zipping along on the Bay Bridge without human intervention, although the surrounding traffic included conventional vehicles being driven by conventional people who were still having to keep their eyes on the road and their hands at 10 and 2.
At the conclusion of the film, van Staa noted with a chuckle that perhaps now Google has finally proven itself useful by providing search capabilities for the disengaged ‘drivers’ using their mobile devices in his film. Given that Googlers work just a few kilometers from where van Staa was holding forth at ICCAD, and might have their own aspirations when it comes to self-driving cars, van Staa’s chuckle was as patently adversarial as his earlier dismissiveness of EDA at an EDA-centric conference.
The message was loud and clear: Bosch is inventing the self-driving car. Through a carefully choreographed combination of engineering disciplines and an artful bringing together of different industries and contributors, the company is gaining mastery of the huge data volumes required for self-driving cars, as well as perfecting safe on-board supply systems, all while keeping an eye on compensation schemes for components that are beginning to exhibit end-of-life issues.
The ICCAD keynoter said Bosch understands how to accomplish mission-aware design and knows how to deal with the difficult electro-thermal issues that arise when ICs are asked to perform perfectly in the hot noisy conditions that exist under the hood. The company also gets that the complexity associated with cars of the future includes analog, digital, mixed/signal challenges, sensors and actuators, peripherals and software, not to mention new materials, plastics and power components.
“If you can close the design gap [related to all of this],” van Staa said, “you’ll be ahead of your competition.” And because nobody else has been able to show Bosch how to close that gap, the company’s been forced to open their own institute, the Robert Bosch Center for Power Electronics.
With their “university” in place, van Staa said his company will now be able to speed up the synthesis of analog circuit models, improve system-level design including the use of virtual prototypes for early software development, and make improvements in verification of software and the configuration of high-level systems.
And with that, the shocking message was complete: Ask not what EDA can do for the cars of the future. Ask instead what Bosch can do for EDA.
Bosch is allowed to test its self-driving cars on California roads because the State legalized the driving of “autonomous vehicles” in 2012. The bill was signed into law by Governor Jerry Brown in a public event on the Google campus in Silicon Valley in October of that year.