Electronic enthusiast original Zhang Ying
From January 9th to 12th, at the Consumer Electronics Show CES 2024 held in Las Vegas, USA, the blessing of AI technology has become a bellwether for exhibitors. Among the more than 4,000 exhibitors, nearly 700 companies in the automotive and mobility technology fields are involved. From chip manufacturers, operating systems to smart car companies, they all showcased their latest technologies and application trends under the AI trend.
Qualcomm, Intel, NVIDIA, Black Sesame Technology, Google, Mercedes-Benz, BMW, etc. have all made a grand appearance, when many car companies and chip manufacturers focus on the development and production of intelligent cockpit and AI on the car, this article will focus on revealing the latest trend of AI technology in the field of intelligent travel and the flagship products of chip manufacturers.
Qualcomm: The integration of cockpit and driving has accelerated, the evolution of Snapdragon digital chassis, and the expansion of the circle of smart car friends
At CES 2024, Qualcomm made two major announcements: first, it demonstrated the device-side artificial intelligence of the Snapdragon XR2+ Gen 2 chip for virtual and mixed reality headsets. The second is Snapdragon's digital chassis portfolio, which will enable the next generation of generative AI, including digital cockpits, connected technologies, connected services, advanced driver assistance and autonomous driving systems.
Pictured: Snapdragon Digital Chassis Concept
On January 10, Qualcomm President and CEO Cristiano Amon told the media that Qualcomm is very excited to enter the field of automotive technology, where cars are increasingly becoming "computers on wheels" that provide drivers and passengers with immersive experiences. In particular, he noted that generative AI will be developed on the device or edge side, not just on the cloud, and that the two will work together. When you walk into the car, the car will become more personalized, which is the change brought about by the application of artificial intelligence technology in the car.
As connected vehicles evolve to central computing and software-defined vehicle architectures, there is a growing demand for highly integrated, high-performance, low-power silicon platforms that can be scaled for multi-level vehicle models. In response to this change, Qualcomm launched the industry's first scalable SoC that supports both digital cockpit and advanced driver assistance systems (ADAS) in 2023, the Snapdragon Ride Flex. Designed to support mixed, critical workloads across heterogeneous compute resources. With pre-integrated hardware, software, and ADAS/autonomous driving (AD) software stack solutions, Snapdragon Ride Flex enables automakers to drive the transition to open, scalable, and integrated architectures in a more convenient and cost-effective way across all vehicle tiers.
图:BMWI7 xDrive60
At CES 2024, the joint booth of BMW and Qualcomm was particularly eye-catching, the BMW i7 is equipped with Qualcomm's Snapdragon Ride and the new cockpit and driving integrated Flex SoC, according to Qualcomm, the Snapdragon Ride platform helps global automakers and Tier 1 suppliers build autonomous driving solutions, up to L3 driver assistance functions such as lane keeping and automatic speed keeping. The Snapdragon Ride Flex SoC enables centralized computing of the digital cockpit, driver assistance, and certain autonomous driving functions on a single system-on-a-chip.
At this year's CES, a new generation of Mercedes-Benz E-Class sedan was exhibited, this car uses the Qualcomm Snapdragon 8259 chip, bringing upgrades and functional innovations to the intelligent cockpit experience, such as the Mercedes-Benz E-Class showcasing the latest in-vehicle technology, including virtual assistants and new audio experiences, MBUX virtual assistants use generative artificial intelligence and 3D graphics technology, through high-resolution images and personalized responses, making the interaction with the vehicle more natural and intuitive.
In addition, among the major international automobile manufacturers, more than a dozen leading automakers and brands such as BMW, General Motors, Volkswagen, and Mercedes-Benz have reached cooperation with Qualcomm to use the Snapdragon Ride platform to create ADAS and AD solutions.
From the perspective of domestic automakers, starting from October 2023, the first batch of cars equipped with Snapdragon 8295 will be released one after another, including Xiaomi SU7, new Mercedes-Benz E-class, NIO ET9, Xpeng X9, Jiyue 01, Zeekrypton 001FR and Zeekrypton 007, Galaxy E8, Leapmotor C10, etc., bringing new paradigms for cockpit functional innovation and experience upgrades, and providing consumers with smart travel and digital life empowered by innovative technology.
From the concept car on the Snapdragon digital chassis at the scene, we see that the edge-side generative AI use cases include four main directions: 1) generative AI-enabled travel guides, which use digital assistants to help plan trips, 2) generative AI-enabled user guides, where users can access an electronic version of the user manual, which provides users with all the details they are looking for, such as how to install car seats, how to change tires, etc., 3) generative AI-enabled predictive maintenance, and 4. Generative AI-powered cockpit control and universal voice queries allow users to focus on the road and drive, allowing the virtual assistant to turn up the cabin temperature or ask the virtual assistant what the weather is like next week. Qualcomm also showcased the Snapdragon digital chassis SoC for two-wheelers for small vehicles and agricultural vehicles.
NVIDIA announced: 2000TOPS NVIDIA DRIVE Thor is the first to be realized on Li Auto
On January 8, NVIDIA announced that Li Auto has selected NVIDIA DRIVE Thor centralized in-vehicle computers to power its next-generation fleet. Great Wall Motors, Zeekr, and Xiaomi have adopted the NVIDIA DRIVE Orin platform to power their intelligent autonomous driving systems.
Automotive architectures are shifting toward centralized computing to enable a high degree of automation and autonomous driving. Drive Thor is Nvidia's next-generation automotive computer that integrates a variety of intelligent functions into a single AI platform that supports functions such as autonomous driving, parking, and driver monitoring. The new Nvidia Drive Thor has a performance of 2000 TOPS, which is better than the existing Nvidia Drive Orin with a performance of 400 TOPS. Li Auto currently uses two Drive Orin processors in its AD Max driver assistance system, and the new AD Max 3.0 upgrade transitions to an AI-centric algorithm architecture. It is reported that Li Auto will be one of the first automakers to adopt the new Nvidia Drive Thor, which is expected to be launched in 2025.
At CES 2024, Wu Xinzhou, vice president of NVIDIA Automotive, said: "The transportation industry is adopting centralized computing to achieve a high degree of automation and autonomous driving. The AI solution of choice for today's smart car manufacturers is NVIDIA DRIVE Orin, and the advanced features and AI performance of their successor, NVIDIA DRIVE Thor, are increasingly being applied to shape future product roadmaps. "
Intel introduces its first SDV SoC family for the automotive market
In addition to Intel's AI-related announcements around its Meteor Lake SoC and its integrated NPU, the automotive market is another market where Intel expects AI to have a significant impact.
On January 10, at the CES2024, Intel shared plans to expand the strategy of AI ubiquity to the automotive market. The plan includes the acquisition of Silicon Mobility, a company specializing in electric vehicle energy management and AI-enhanced software-driven digital cockpits. The acquisition is intended to strengthen Intel's ability to deliver efficient energy management solutions for electric vehicles and AI-driven ecosystems.
Figure: Jack Weast introduces the automotive software-defined vehicle SoC family
Also, during CES 2024, Intel unveiled a new family of AI-enhanced software-defined vehicle (SDV) SoCs. The details of this product are the scalable SoC that includes up to 12 cores and a power range of 12 to 45 W, depending on the cockpit level and EV power requirements. They also feature integrated Arc Xe integrated graphics that support up to 4 Indept displays with resolutions up to 8K. The automotive industry has very stringent security standards, so the new Intel SDV SoC family is AEC-Q100 qualified.
It is reported that the chips leverage Intel's expertise in artificial intelligence, such as generative AI and camera-based driver/passenger monitoring. Geely's Zeekr brand has been announced as the first OEM to adopt these new Intel SoCs in the automotive sector.
Intel's Mobileye announced the launch of the Eye Q7H chip, which is manufactured on a 5nm process with 67TOPS computing power and 60W power consumption, and is scheduled to achieve mass production in 2027.
Challenging cross-domain computing, Black Sesame Intelligent announced the C1200 series mass production model
On January 9, Black Sesame Intelligence, an important manufacturer of intelligent vehicle computing chips in China, brought its automotive-grade high-performance autonomous driving chip Huashan series A1000 mass production ecology and intelligent vehicle cross-domain computing chip Wudang C1000 family to CES exhibition.
The A1000 series chips have supported a variety of applications, including integrated driving and parking, city navigation, valet parking, 3D surround view, vehicle-road coordination, driver monitoring system, etc. At present, the A1000 series is in a state of full mass production and delivery, and has been adopted by FAW Group, Dongfeng, Geely, JAC Group and other car companies.
Pictured: Black Sesame Intelligence at CES 2024
An important highlight of Black Sesame Intelligence at CES 2024 is that it has disclosed for the first time the mass production model of the C1200 family of cross-domain computing chip series, with two models, one is the C1236 that supports the integration of single-chip NOA driving and parking, and the other is the C1296 that supports multi-domain integration. It is reported that the C1200 series can integrate multi-domain functions with a single chip, including cabin and parking integration, 6V BEV perception display, Vehicle data exchange display, hard isolation and GPU rendering display.
Black Sesame Intelligence, together with JOYNEXT and its software subsidiary, JOYNEXT, showcased a series of software exhibits including nCCU - Central Computing Unit products, AUTOSAR product solutions, and functional domain software solutions based on the Wudang series C1200 family intelligent vehicle cross-domain computing chip platform.
Google brings new Android Auto features to more car brands
If you're using Android Auto in your vehicle or driving a vehicle with built-in Android software, Google has introduced some new features at CES 2024. First of all, Android Auto is making an upgrade to electric vehicles, and range-anxious motorists should appreciate this feature. Google Maps will provide an estimate of the vehicle's battery level when the driver arrives at their destination, and highlight charging locations along the way.
Image from Google's official website
In addition, for a more low-profile and easy-to-use upgrade, Google Maps on Android Auto can now load trips you plan on your phone on cars that have Android Auto installed.
Also, for cars with Google Android built-in, the Weather Channel app is now also available, where users can view hourly weather forecasts that follow the driver, receive severe weather alerts, and see the "Trip View" radar on the dashboard.
When it comes to entertainment, cars with Android built-in can now add multiple video streaming services and watch featured shows from PBS Kids and Crunchyroll directly from the dashboard screen.
Google's official website announced that these EV-specific features will be coming to the Ford Mustang Mach-E and F-150 Lightning "in the coming months," with more EVs with Android Auto to follow.
Currently, Volvo, GMC, Cadillac, Honda, Polestar, Renault, and Chevrolet vehicles have Google built-in. Google says that Nissan, Ford and Lincoln models will be launched this year, and Porsche plans to launch them in the near future.
brief summary
At CES 2024, the intelligent cockpit is evolving into an emotional space centered on user experience, and Qualcomm, Intel, and Nvidia have added AI algorithms to chips or integrated multiple functions into AI platforms, all of which have catalyzed the formation of a new trend of smart cockpits during the continuous increase in the penetration rate of smart cars.
In addition, with the implementation of central computing platforms from manufacturers such as NVIDIA, Qualcomm, and Black Sesame, 2024 will be the first year of cross-domain (regional + central computing). For both OEMs and suppliers, new market opportunities have arrived.