Audi shows off next-gen in-car system that can predict the future

At the GPU Technology Conference, Audi has shown off Urban Intelligent Assist, the German automaker’s vision for the future of urban driving. Urban Intelligent Assist (UIA) is a suite of hardware and software tools that, among other things, predicts the future and provides natural, landmark-based GPS navigation.

According to Audi, more than 80% of Americans now live in an urbanized area — a 12% increase over the last decade. As you have probably noticed, this means it’s now harder than ever to drive in cities — and indeed, most road traffic accidents occur on urban areas. 


Increased car density obviously doesn’t help matters, but Audi also found that driver frustration and distraction also play a huge role. With features such as lane keep assist and blind spot sensors, cars already go some way to reducing frustration and distraction — but Audi thinks they should do more; a lot more.

Which brings us neatly onto Urban Intelligent Assist, a hardware and software suite which includes Predictive Traffic, Smart Parking, Naturalistic Guidance, and Driver Attention Guard. 

All of these features are rather awesome, but Predictive Traffic is probably the coolest. By analyzing past traffic records, current road conditions, and any upcoming events — such as football games or Black Friday — Predictive Traffic generates the best route to your destination. The slide shown at the conference also mentions that Predictive Traffic will predict your next destination, based on your driving history.

Dovetailing into Predictive Traffic is Naturalistic Guidance, which is basically GPS navigation with landmarks thrown into the mix. Instead of “in 100 yards turn left,” you might hear “take a left past the Hilton.” Once you near your destination, Smart Parking will guide you to the nearest open parking spot — and show you how much each parking spot costs. This feature only works in cities with smart parking bays/meters that publicly announce their status, such as San Francisco and New York, but more cities will come online in the next few years. Once you park, Audi’s Seamless Navigation will transmit the remainder of your route to a special app on your smartphone — an app that also acts as a standalone trip planner, telling you when you must depart by to make it to your destination on time.

Finally there is Driver Attention Guard, which will trigger various autonomous-driving features if you’re not paying attention to the road. By monitoring you with multiple cameras and detecting when your attention wanders — looking away from the road, your grip on the wheel loosens — Attention Guard will use adaptive cruise control (and presumably other features such as lane keep assist) to prevent accidents. Your car will also try to regain your attention, probably through some combination of flashing lights, beeps, or a vibrating steering wheel.

At this point you might be wondering why Audi unveiled UIA at the GPU Technology Conference (GTC), which is where AMD and Nvidia usually show off their latest super-mega graphics cards or scarily-real real-time graphics demos. As far as we can tell, Audi chose the GTC because the actual UIA in-car unit will be powered by a Tegra chip, and Nvidia asked Audi to come along and show off the kind of in-car solutions that Tegra is powering. Nvidia’s Tegra 3 SoC — or Visual Computing Module, to give its in-car name —  is already being used in the Tesla Model S, and automakers such as GM (Chevy) and Audi are planning to roll out VCM-based infotainment units.

As for when we might actually see Urban Intelligent Assist on the road, we are probably looking at a timeline of a few years. Audi says it’s already testing all of these features in San Francisco, and most of the constituent parts — adaptive cruise control, landmark GPS — are already mature technologies. (See: Audi shows off self-driving car at CES – no sign of Google.) It’s just a matter of wrapping it all into a neat package, which is a much faster process now that automakers can target a well-known platform such as Tegra, instead of messing around with proprietary, bespoke systems.

Extremetech