Global platforms require a greater emphasis on systems-engineering and requirements management, according to Marc Halpern, Research VP at the Stamford, CT-based analyst firm Gartner Inc. “You need, through your systems models, to be able to identify the critical design parameters, and how those interact with other parameters across subsystems and parts, in order to be able to design a truly global car.” Along with the systems engineering, he added, is the role of requirements management in addressing environmental, economic, safety, or ergonomics targets.
Jon Friedman, a former automotive engineer and current industry marketing manager for aerospace, defense, and the automotive industry at the Natick, MA-based mathematical computing software company MathWorks (www.mathworks.com), says executable model-based design assists engineers looking to satisfy those different preferences. The tools allow teams to create a generic algorithm that can be tuned for local markets, and then verify that the given embedded system functions properly.
If a platform’s suspension works well everywhere except in Brazilian markets, where the roads are rockier, or if an engine controller needs to be fine-tuned, engineers can explore what Friedman calls the “outer performance envelope” of components, and then work to determine the authority of the tuning component to make sure that they can meet a wide variety of performance goals.
“For a suspension, an engineer might want to determine how much of that suspension system, that shock absorber, is tunable. Can it be both soupy and tight? Those characteristics can be modeled inside of a mathematical modeling tool and can be analyzed at a system level to determine whether overall the vehicle behavior on the edges of the performance envelope will meet the local market requirements,” said Friedman.
Hickok calls analytical models the “life blood” of the early design process, and a way to ensure that the right content is specified for the architecture before hardware is built.
If a major critical design issue is found in a created hardware component, like Hickok’s suspensions, it is often too late to implement a change in the math to impact vehicle builds and still meet the start of production without significant risk.
Hickok emphasized that models are often most critical as products are slowly rolled out. “As we launch in one region and the remaining programs follow over some pre-determined cadence, you may not see hardware representative of the ‘India’ product, for example, until long after you’re nearing production for the lead program, so changes found in hardware for India will be challenged to find their way back into the architecture without putting all program variants at risk.”
Once engineers move from design to implementation, models made from analytical tools like MathWorks’ Simulink can automatically generate code: both code to run prototype and code for production. “A lot of the cars today run production controllers with code that was automatically generated from mathematical models,” said Friedman.
The Importance of Software
Halpern believes that on-board software has become a game-changer for global cars, because of its ease in allowing variation. GPS and vehicle security features, for instance, enable communications and connectivity regardless of location. Software, he said, enhances variation of vehicle function depending on local market preferences.
“Imagine on your speedometer or tachometer, if you want text in French versus Italian or English. It’s much easier to enable text or even voice preferences via software than it would be to have physical parts with printing in different languages.”
The auto industry is also increasingly working to standardize a vehicle’s software infrastructure, bringing more plug-and-play ideas into the way embedded software is deployed in standard features, including headlight control and windshield wipers. AUTOSAR (Automotive Open System Architecture), an open and standardized automotive software architecture developed by automobile manufacturers, suppliers, and tool developers, aims to create one common standard for electronics and software in vehicles.
“[Engineers] have to really rethink the growing role of software as part of a car platform, and whatever they’re designing, how much of the function of what they typically design is going to be in the future addressed by software,” said Halpern, adding that the Chevy Volt, according to media reports, contains at least 10 million lines of code.