Car stuff (public number: chedongxi)
The author | Hao Han
Edit the | Juice
Car East March 23 news, the chief intellectual officer conference and machine heart AI technology annual meeting opened this afternoon, it is understood that the conference organizers invited more than ten industry insiders in the fields of vehicles, chips, automatic driving, unmanned logistics and other fields to participate in the interview, share their latest views on the field of automotive autonomous driving and pre-judgment of trends in the field.
The conference will focus on the themes of intelligent car 3.0 era, intelligent driving data, bicycle intelligence, cloud edge end collaboration, and the commercialization of Robotaxi in China, and will also invite guests to discuss how to mature the commercialization of chips and autonomous driving in the era of large computing power.
As the third zhixing official guest of the whole event, Wang Ping, CEO of Cambrian Xingge, shared the theme of "Bicycle Intelligent Breakthrough, Cloud Edge End Vehicle Collaboration".
Cambrian Xingge is a subsidiary founded by Cambrian focusing on autonomous driving chips, which has received investment from SAIC, Weilai Automobile, CATL and other enterprises.
▲Cambrian Xingge EXECUTIVE PRESIDENT Wang Ping
First, in the next five years, L2+ will be rapidly popularized Large-scale mass production of L4 still needs time
Wang Ping said that the current development of autonomous driving is like a mountain road full of opportunities but rugged.
In the next five years, there will be three major trends in automatic driving, one is that the L2+ level automatic driving system will be rapidly popularized and exist for a long time; the other is that the L4 level automatic driving in restricted scenarios will gradually land, but there is still a long way to go before large-scale mass production; the third is that the vehicle-road cloud collaboration realizes a big data closed loop, which can promote the continuous upgrading of the driving experience.
At present, the scale of the automatic driving system is still facing multiple challenges, the first is the bicycle intelligent driving system.
▲The scale of the automatic driving system faces multiple challenges
Wang Ping believes that at present, the processing capacity of a single SOC is generally insufficient, requiring 2 or even multiple pieces cascading, the system complexity is high, and mass production is also difficult. In addition, high power consumption, high cost, and low proportion of domestic chips are also problems faced by the current bicycle intelligent driving system.
In addition, there are also three challenges in the coordination of vehicle-road cloud, one is that the massive data closed loop requires large-scale AI clusters, and the cost is relatively high, the second is that data security and privacy protection are challenged, and the third is that the personalized needs of car owners are difficult to meet.
Wang Ping believes that "universal openness" and "large computing power" will become the two major trends of intelligent driving chips.
Second, this year and next two years to launch two blockbuster chips full coverage of L2+ and L4 scenes
Cambrian Xingge, as an in-vehicle chip company, is working with ecological partners to develop smart driving chips that meet the needs of various levels of market demand.
Wang Ping said that Cambrian Xingge will launch products covering different scenarios and different levels of automatic driving. This year and next year, Cambrian Will launch two blockbuster chips, one for the L4 market, which can support end-of-the-car training, and the other for the L2+ market.
It is reported that this high-end intelligent driving chip for the L4 market will use a 7nm process, WITH AI computing power exceeding 400TOPS, CPU maximum computing power exceeding 300K+DMIPs, and supporting car-side training.
▲High-end intelligent driving chip series for the L4 market
The entry-level chip AI computing power of the L2+ market reaches 15TOPS, and a single SOC can realize the function of L2+ berth integration, with low power consumption and other characteristics, which will promote the popularization of automatic driving to 5 to 100,000 yuan of entry-level vehicles.
In addition, Xingge Technology will also cooperate with parent company Cambrian to launch a cloud-edge end car solution, which will be based on unified platform-level basic software, processor architecture and instruction set.