Audi: Machine Learning to Give Cars Superhuman Capabilities

by Brian Caulfield

Machine learning will give cars the ability to analyze and learn from hundreds of thousands, even millions, of driving situations to learn better than any human being can, Audi execs told the press at CES 2016 Wednesday.

“In one year a 16-year-old driver might encounter 1,000 situations with a four-way stop,” said Peter Steiner, managing director of Audi Electronics Venture. “But our systems will learn from hundreds of thousands, even millions, of such situations that can be stored, analyzed and improved from, so these cars can learn even better than a human being can.”

Audi is continuing to work toward fully autonomous vehicles — it sent a self-piloted A7 sedan from San Francisco to Las Vegas in time for last year’s CES — even as it continues to introduce increasingly sophisticated advanced driver assistance features that have seen a strong uptake among Audi buyers.

ces audi cockpit web
Digital cockpits were a highlight for Audi at CES 2016.

Audi continues to work closely with NVIDIA, incorporating our Tegra processors into its zFAS driver assistance control unit and MIB infotainment system.

“Due to our close collaboration with NVIDIA, Audi has the ability to integrate technology quickly and move at the same innovation cycle as the consumer electronics industry,” said Ricky Hudi, the carmaker’s head of Electrics/Electronics.

Audi is working toward connecting advanced automotive sensors and the high-definition, cloud-connected mapping capabilities of HERE — which it co-owns with BMW and Daimler-Benz — to create a wealth of data its deep-learning systems can use to create ever smarter driver-assistance systems.

“We will push the technology of artificial intelligence, of machine learning, to get the same recognition rate — or even better than a human being,” Hudi said.

“For me this is the most disruptive technology — machine learning — in what you can discover now, and not just in the automotive industry,” he added.