REM uses the cameras that are already on the front of many vehicles to pick out landmarks and other road details. The tech uses a tiny amount of bandwidth each kilometer to send info to OnStar, but GM could have maps that are always up to date by multiplying the collection across a whole range of vehicles. Mobileye believes the data could be accurate within about four inches. Driverless cars could then use this knowledge in combination with their other sensors to navigate themselves more safely. "Creating and updating maps using on-board camera technology supplies the missing link between on-board sensing and the requirement for full redundancy to enable safe autonomous driving," said Amnon Shashua, co-founder, chairman and chief technical officer of Mobileye.
GM currently has five vehicles testing REM and plans to grow the fleet to about 20 soon, spokesman Alan Adler told Autoblog. To alleviate privacy concerns, OnStar would collect the data in aggregate and keep vehicles anonymous, he said.
"We are planning to explore the integration of REM into existing GM program launches sometime later this year," product boss Mark Reuss said in the tech's announcement.
Like much of the auto industry, GM will take driverless technology development very seriously in the coming years. For example, the company will soon test an autonomous carsharing service for employees at its Michigan Proving Grounds, and it will invest $500 million in Lyft for a similar venture.
Road Experience Management targets accuracy to enable fully autonomous driving
DETROIT – General Motors is exploring a new mapping technology from Mobileye that could use crowd-sourced real-time data collected by OnStar for precise localization and high-definition lane data that supports fully autonomous driving.
GM is Mobileye's largest customer and is engaged with Mobileye Road Experience Management (REM) to create a continuously updated road map. Mobileye's existing EyeQ processing platforms extract landmarks and roadway information at extremely low bandwidth of about 10kb per kilometer.
"Cameras are the most natural sensors for creating maps because they are already available in most new car models as part of the trend toward growing driver assistance deployment," said Amnon Shashua, co-founder, chairman and chief technical officer of Mobileye. "Creating and updating maps using on-board camera technology supplies the missing link between on-board sensing and the requirement for full redundancy to enable safe autonomous driving."
Mobileye expects to eventually achieve localization at an accuracy of about 10 centimeters compared with today's GPS, which can pinpoint to about 10 meters.
OnStar serves more than 7 million customers with a 24/7 human touch, in the U.S., Canada, China, Mexico, Europe and Brazil. OnStar recorded its 1 billionth customer interaction in 2015.
GM, which sold about 3.3 million vehicles in the U.S. in 2015, uses Mobileye software on cameras that support existing safety technologies like forward collision alert. GM has collaborated on camera technology with Mobileye for a decade beginning with lane departure warning systems launched in 2007.
"GM is committed to bringing semi-autonomous and fully autonomous vehicles to our customers, and this technology will be a critical enabler to getting us there," said Mark Reuss, GM executive vice president of Product Development, Purchasing and Supply Chain. "We are planning to explore the integration of REM into existing GM program launches sometime later this year."
General Motors Co. (NYSE:GM, TSX: GMM) and its partners produce vehicles in 30 countries, and the company has leadership positions in the world's largest and fastest-growing automotive markets. GM, its subsidiaries and joint venture entities sell vehicles under the Chevrolet, Cadillac, Baojun, Buick, GMC, Holden, Jiefang, Opel, Vauxhall and Wuling brands. More information on the company and its subsidiaries, including OnStar, a global leader in vehicle safety, security and information services, can be found at http://www.gm.com