During the CES 2022 keynote, Mobileye, an Intel company, introduced its EyeQ Ultra, the most advanced and high-performing SoC for autonomous vehicles (AV).
At only 176 TOPS, EyeQ Ultra is the industry’s leanest autonomous vehicle (AV) chip, maximizing both efficacy and efficiency. This well-designed SoC draws on seven generations of proven EyeQ architecture to give exactly the power and performance required for autonomous vehicles, which will almost certainly be all-electric.
The EyeQ Ultra SoC’s first silicon is expected at the end of 2023, with full automotive-grade production beginning in 2025.
“Consumer AV is the end game for the industry,” said Prof. Amnon Shashua, Mobileye president and chief executive officer. “By developing the entire self-driving solution – from hardware and software to mapping and service models – Mobileye has a unique perspective into the exact requirements for the self-driving system that enables us to reach the performance-and-cost optimization that will make consumer AVs a reality.”
After building an AV to understand exactly what a self-driving car needs to run at a very high mean time between failures, Mobileye produced the EyeQ Ultra. This method achieves the best performance balance between accelerators and general-purpose processors while remaining inside a very tight power-performance envelope.
The EyeQ Ultra is a significant step forward in the evolution of the EyeQ family of SoCs, combining the performance of ten EyeQ5s into a single device. EyeQ Ultra, which is based on 5nm manufacturing technology, can handle all of the needs and applications of Level 4 (L4) autonomous driving without the power consumption and expenses associated with combining numerous SoCs.
EyeQ Ultra, very much like its predecessors, was designed in partnership with Mobileye software to achieve exceptional power efficiency while sacrificing minimal performance.
EyeQ Ultra makes use of four different types of patented accelerators, each designed for a distinct task. These accelerators are combined with more CPU cores, ISPs, and GPUs to create a highly efficient solution capable of processing data from two sensing subsystems – one camera-only and the other radar and lidar combined – as well as the vehicle’s central computing system, high-definition map, and driving policy software.
The EyeQ Ultra is far more efficient than other AV solutions, giving the required performance and price-point for consumer-level AVs at a mere 176 TOPS. EyeQ Ultra unlocks the AV potential for safer roads and less congestion for consumers by optimizing for efficiency.
EyeQ Ultra is being released at the same time as two new EyeQ SoCs for ADAS, the EyeQ6L and EyeQ6H, and comes after Mobileye shipped its 100 millionth EyeQ SoC late last year.
Mobileye’s EyeQ, first presented in 2004, revolutionized the ADAS market by demonstrating that low-cost video sensors processed by Mobileye’s purpose-built technology could avoid and mitigate crashes.
EyeQ’s innovation aided in the accessibility of roadway safety technology, providing features such as forward-collision warning, lane departure warning, and blind-spot identification to millions of drivers all over the world.
EyeQ Ultra is built on Mobileye’s well-proven EyeQ architecture. EyeQ Ultra, like the EyeQ line of SoCs before it, addresses a need in the automobile market by making consumer AVs accessible.
EyeQ Ultra will be informed by Mobileye’s Road Experience Management (REM) mapping technology as a member of the EyeQ family. REM collects packages of road data from millions of vehicles on the road already equipped with Mobileye to generate the Mobileye Roadbook, which is accessed via the cloud to offer real-time information on the drivable roads ahead.
The EyeQ6L is the successor to the EyeQ4 SoC and comes in packaging that is just 55% the size of the EyeQ4. For highly efficient entry and premium (L2) ADAS, our one-box windshield solution delivers additional deep-learning TOPS at ultra-low power. It started sampling last year and expects to begin production in the middle of 2023.
With full surround, the EyeQ6H will provide premium ADAS or partial AV capabilities. In terms of computational capacity, it is equivalent to two EyeQ5 SoCs, but it also supports visualization and performs better under intensive artificial intelligence workloads.
All ADAS L2+ functions, multi-camera processing (including parking cameras), and third-party apps such as parking visualization and driver monitoring will be available through this unified system. This EyeQ family’s most advanced ADAS SoC will start sampling this year and go into production by the end of 2024.
Both EyeQ6 SoCs will be built using a 7nm process.