As self-driving startups adjust their strategies (e.g., from Robotaxi to trucks), are busy with the next round of funding (there are fewer primary market opportunities and more secondary market opportunities, such as listing through SPAC), and announce their own plans for vehicle-grade, mass-production products, veterans of the automotive industry are overtaking in curves.
From Robotaxi to privately available self-driving cars, that's the answer given by Mobileye and General Motors during this year's CES show, and the timing set by the two companies is almost surprisingly the same: around 2025.
Among them, GM Chairman Mary Barra's goal is to "become the fastest company to market for personal retail self-driving cars." Behind this is the technical support of self-driving subsidiary Cruise and a personal self-driving electric concept car (InnerSpace) unveiled by Cadillac at this year's CES.
Mobileye announced that it has further deepened its strategic partnership with Geely's Extreme Krypton brand, and the two sides plan to jointly launch a consumer-grade intelligent driving electric vehicle with L4-level capabilities as soon as 2024, which can be sold to individual users.
Before that, though, both companies will have to go through a transition to technological upgrades.
One
For GM, Ultra Cruise will be a transitional goal.
The highly anticipated Cadillac Celestiq, a pure electric coupe, will go on sale in 2023, becoming the brand's first new car to feature ultra Cruise. In terms of functional settings, this is an upgraded version of the current GM Super Cruise, allowing door-to-door "hands-off" self-driving.
In terms of powertrain, the car will be equipped with GM's Ultium power battery pack, with a battery capacity of between 50-200 kWh, a range of more than 644 km, an all-wheel drive and four-wheel steering technology, and possibly 350 kW fast charging technology.
Unlike Super Cruise, Ultra Cruise is equipped with a combination of camera, radar and lidar sensors, supports automatic and on-demand lane changes, and the HD map covers more than 2 million miles (3.2 million kilometers), nearly 10 times the former, and allows urban streets, block roads and rural road conditions to achieve 95% driving scene coverage.
For Super Cruise, this L2-level driver assistance system is not "leading" the industry in terms of technology and experience. What really makes this feature famous is simply that there is an additional driver monitoring system than Tesla to circumvent possible system abuse by drivers.
According to the plan, Super Cruise and Ultra Cruise will coexist as two different product lines, of which the former is used for low-end and low-end models, while the latter is only available for high-end models.
From the information disclosed so far, Ultra Cruise is based on Qualcomm's 5nm, scalable computing architecture for future software OTAs through the Ultifi software platform and vehicle intelligence platform. At the same time, on the basis of Super Cruise, support for V2X, close-up object collision avoidance, residential memory parking, etc. has been added.
At the same time, the new intelligent diagnostic and self-learning system will automatically identify scenarios where Ultra Cruise needs to be upgraded, trigger data recording on vehicles equipped with the service, and process it through GM's back-end data ecosystem to continuously improve the system.
In addition, the entire software platform will be built on top of GM's VIP electronic hardware architecture, providing powerful data processing capabilities that can handle up to 4.5TB of data per hour, support high-speed in-vehicle Ethernet communication, and be fully prepared for technologies such as V2X intelligent transportation.
From the current public information, Ultra Cruise may be available to users in the form of a monthly subscription fee. In terms of the corresponding functional level, Ultra Cruise is still designed as an L2 level automatic driving assistance system. The features may be more powerful and can be used on more roads, but still require the driver to maintain attention at all times.
According to public information, the Ultra Cruise system was developed by GM's research and development team, and Mary Barra's goal is to "graft" Cruise's self-driving capabilities to GM. That's why Dan Ammann abruptly "left" from crusie's CEO position, which was originally hoped to push Cruise to go public independently.
Obviously, from the perspective of General Motors, at this stage, it is necessary to adopt a "two-way" breakout model, and relying solely on Robotaxi is not the only choice for car companies. In this regard, the same goes for Volvo Cars.
According to the plan, Volvo Cars will release a new generation of pure electric large-scale SUVs later this year, and for the first time launch the highway L4 level autonomous driving technology. The technology, called Ride Pilot, will first be rolled out in California and begin driving road testing in China in the middle of this year.
The system is based on Volvo Cars' partnership with Zenseact (a software subsidiary) and Luminar (which provides lidar) to use 1 Luminar Iris lidar, supplemented by 8 cameras, 16 ultrasonic sensors and 5 millimeter-wave radars.
Martin Kristensson, head of AD and Mobility at Volvo Cars, said, "When we say autonomous driving, we mean autonomous driving. So it would be a feature driven by the car so that you don't need to put your hands on the steering wheel or keep your eyes on the road. ”
Two
For Mobileye, it's mostly a game of time and cost.
In Mobileye's view, the relationship between Robotaxi and consumer-level autonomous driving is not conflicting, but synergistic, and it is important to establish this relationship. From this point of view, GM is also seeking a synergistic breakthrough.
"The reason we are now able to afford consumer-grade autonomous driving capabilities is because in our Robotaxi solution, there are very strict design constraints on the cost of the system." Erez Dagan, executive vice president of product and strategy at Mobileye, said.
On January 4, Mobileye unveiled its most advanced and powerful EyeQ Ultra system integration chip for autonomous driving to date at CES, optimizing energy efficiency at 176 TOPS and providing performance and power consumption for all autonomous vehicles, especially all-electric autonomous vehicles.
The EyeQ Ultra's performance is equivalent to the sum of the performance of 10 EyeQ 5s, and with the help of a 5nm process, it can meet all the needs and application scenarios of L4 autonomous driving, while avoiding the additional energy consumption and cost of combining multiple system integration chips.
Like its predecessor, the EyeQ Ultra was co-designed with Mobileye software, enabling the chip to achieve extremely high energy efficiency ratios without sacrificing performance. A set of four proprietary accelerators is used, each designed for a specific task.
In addition, the solution is capable of simultaneously processing input data from two sensing subsystems (including a pure camera subsystem and a subsystem that integrates radar and lidar) as well as the vehicle's central computing system, high-definition maps, and driving decision software, providing the performance required for consumer-grade autonomous vehicles at a reasonable cost.
According to the current plan, EyeQ Ultra is expected to start delivery to users by the end of 2023 and fully achieve vehicle-grade mass production in 2025. This means that the L4 autonomous driving system equipped with the EyeQ Ultra solution will be sold on a large scale through the private market until 2025.
Mobileye and Krypton's cooperation discloses that the L4 level of the two parties is for private retail autonomous driving models, which are based on six EyeQ5 system chips as the hardware basis, to achieve the efficient integration of Kr proprietary software technology with technologies such as true redundant sensing solutions, responsibility-sensitive security systems (RSS) and road network acquisition management (REM).
At the same time, Mobileye announced that it has collected 200 petabytes of public road real driving data, which means that even "long tail" events in rare conditions and scenarios can still output thousands of decision results in a matter of seconds.
As we all know, cost is the key to whether L4 autonomous driving can be oriented to the private consumer market. Previously, Mobileye expected that the cost of L4-level hardware and software solutions would be expected to drop below $5,000 by 2025.
In addition to the accumulation of chip and vision technology, it will also rely on the company's self-research on FMCW lidar and 4D millimeter wave radar.
"Consumer-grade autonomous vehicles are the endgame game for the industry." In mobileye's view, Amnon Shashua, CEO of Mobileye, by developing entire autonomous driving solutions (from hardware and software, from maps to end-to-end models) can achieve performance and cost optimization, and the goal will gradually become a reality.