A global automotive supplier thinks carmakers are wasting too much time developing software that runs autonomous technology.
In working with clients, engineers at Elektrobit, a subsidiary of Continental, say automakers have become adept at creating standalone features like adaptive cruise control and active lane-keeping assist. But these features aren’t naturally designed to collaborate with each other, and as automakers ramp up development of autonomous driving, that’s exactly what they’ll need to do.
So Elektrobit has designed a new software architecture that’s intended to eliminate much of the legwork in making those connections.
Rather than automakers each developing their own proprietary connections between sensors and features, Elektrobit says its new software platform, dubbed “EB Robinos” standardizes integration of driver-assistance functions into one system. Company officials believe it could shorten development time for automated-driving systems by as much as 30 percent.
That remains to be seen, of course. But in an industry where traditional automakers, suppliers, and Silicon Valley tech companies all are clawing for any advantage in bringing self-driving technologies to the market, EB Robinos might represent a way for the players to move beyond the basics of what’s essentially developmental busywork and concentrate their efforts on ways to differentiate their semi-autonomous or autonomous offerings.
“Creating, coordinating, and combining these individual parts presents significant challenges,” said Phil Magney, an automotive functional safety analyst at Vision Systems Intelligence. “EB Robinos consolidates many elements of the tool chain into a single open platform that can speed development of any autonomous-driving platform.”
Elektrobit announced the creation of Robinos in late June, and automakers and other suppliers are scheduled to start using it this month, according to the company. A more comprehensive version that allows for path planning and grid fusion is slated to arrive by the end of September.
Though it won’t say who the first customers are for Robinos, Elektrobit has worked heavily in the past with OEMs such as Audi and BMW, providing infotainment, driver-assistance systems, and connected-car software. Earlier this year, Elektrobit provided software that assisted in motion prediction and path planning for an autonomous-shuttle project in Europe called WEpods. Continental purchased Elektrobit’s automotive software division last year for $665 million, according to a report by Bloomberg.
In simpler times, it was easy for automakers to design driver-assist features themselves. An adaptive cruise-control system, for instance, had four components that needed to interact—a speed control, radar, a module to help understand which car it should follow, and a user interface. For autonomous driving, those components might also need to interact with cameras, LiDAR, ultrasonic sensors, mapping, cloud data, GPS, and more. The complexity has mushroomed.
Even so, there’s a need to keep development simple. Manuela Papadopol, global marketing director for Elektrobit, likens it to the need for the basic driving experience, functionally speaking, to be the same no matter which car a motorist drives.
“A vehicle has to work the same way, whether I get behind the wheel of a Ford in Europe or here, it has to work the same,” she said. “Some applications, like traffic, might need to be adapted. But the basic functionality is the same, and that’s what autonomous software should look like. That’s what Robinos does. It’s a fundamental operating system that’s good for developing features and functionality.”
Robinos allows developers to tailor a system to specific needs, with the potential for as many as 24 different software modules utilized at one time, according to Jürgen Ludwig, chief of business development for Elektrobit. And it permits them to swap out parts while minimizing the need to revalidate each component after a single change.
More broadly, it might also allow for more competition in the race toward autonomy, allowing carmakers and suppliers who are behind in development to play catch-up.
“Some compares we aren’t even thinking about as ‘autonomous car companies’ that might want to enter the space, they can use it for prototyping,” Papadopol said. “But certainly for automakers, this could be a differentiator. If you an save 30 percent of your development time, that’s huge if you think about it. Everybody is racing to be the first autonomous car on the road, and it’s just not about that, but it’s about how safe it will be and how easy it is for customers to use. From a consumer-adoption standpoint, I think carmakers will benefit, but ultimately, so will the drivers.” Although we suppose they may not be called “drivers” for long.
from Car and Driver Blog http://ift.tt/2aPsacB
via IFTTT
0 comments:
Post a Comment