Mobileye, Nvidia (and Others) Spar over Cars

LAS VEGAS—The field of autonomous driving technologies revealed its substantial expansion at this year’s International Consumer Electronics Show (CES).

In the autonomous driving chip segment alone, aside from incumbents such as Nvidia, Mobileye, NXP and Texas Instruments, a host of new players — Ceva, an IP vendor, Intel and Qualcomm — are now also breaching the market. Car OEMs are welcoming these newcomers, observed Egil Juliussen, director of research, infotainment & ADAS at IHS Automotive, during CES. “The field has suddenly gotten a lot more crowded.”

Fog of war?
By now, the investment and media communities are thoroughly tuned in to the technologies that enable autonomous cars — sensing, cameras, radars and lidars, mapping, algorithms, deep network (or not deep network), artificial intelligence, etc.

But it still remains unclear — to most of them — where all these technologies will eventually end up in autonomous car designs, let alone who will be the winners and losers in this battle.

Mobileye CEO

Mobileye CEO

Mobileye co-founder and Chief Technology Officer Amnon Shashua noted that he first thought competitors were deliberately fomenting misinformation [about all these technologies] to create “the fog of war.” But now he realizes that “people are simply very confused. They just don’t know.”

Nvidia with its “deep learning” pitch, and Mobileye with a mapping technology unveiled at CES, were the two clear stars at the show, each fiercely jockeying for position in the race for advanced driver assistance systems (ADAS) and autonomous driving.

Mobileye didn’t hesitate to publicly knock Nvidia during Mobileye’s press conference. CEO Shashua said, “What I saw in Nvidia’s announcement was a liquid cooled super computer with 250 watts of power that costs $10,000. I think it’s not relevant to our space.”

Ceva’s CEO Gideon Wertheizer described the public spat between the two as “intended for investment play.” Indeed, Mobileye’s stock dipped almost 10 percent earlier last week following announcements from Nvidia, but climbed again after its CES press conference.

Crowd-sourced mapping
But Mobileye’s announcement did feature certain substantive technology components. Wertheizer, in a one-on-one with EE Times, singled out Mobileye’s new mapping technology development — called Road Experience Management (REM) — potentially as “most dangerous” to rival chip suppliers and Tier Ones such as NXP, Bosch and Denso.

According to Mobileye, REM creates “crowd-sourced real-time data” for precise localization and high-definition lane data — an essential layer of information to support fully autonomous driving. 

How Mobileye's REM works in the backend (Source: Mobileye)

How Mobileye's REM works in the backend (Source: Mobileye)

The technology is based on software running on Mobileye's EyeQ processing chips. It extracts landmarks and roadway information at extremely low bandwidths — roughly10 kilobit (Kb) per kilometer of driving. (Google, in contrast, does its localization and high-definition mapping at Gigabit per kilometer.) Mobileye explained that backend software running on the cloud integrates the segments of data sent by all vehicles with the on-board software into a global map.

Mobileye’s visual interpretation scheme (which helps to compress data) will supposedly enable automakers to create their own “road book.”

To read the rest of this article, visit EBN sister site EE Times.


0 comments on “Mobileye, Nvidia (and Others) Spar over Cars

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.