What do industry leaders think about autonomous driving? After 2025, it will mature, and different technical routes will have their own strengths

Car stuff (public number: chedongxi)

Author | Alice

Edit | Juice

On March 23, the Chief Intellectual Officer Conference and the Machine Heart AI Technology Annual Conference opened, and a roundtable debate was held on how to mature the commercialization of autonomous driving.

Zhou Xin, co-founder and chief product officer of Yishi Technology, Hao Jianan, co-founder and chief architecture officer of Tucson Future, Dong Jian, co-founder and vice president of software algorithms of Hongjing Intelligent Driving, and Dai Zhen, vice president of Heduo Technology, participated in the roundtable debate.

▲ Introduction of executives participating in the roundtable debate

Here's a distillation of the core elements of this roundtable debate:

1. The full-scenario autonomous driving technology route will accelerate the maturity of technology.

Zhou Xin, co-founder and chief product officer of Yishi Technology, believes that Robotaxi is the endgame scenario of autonomous driving.

But the process of developing Robotaxi is very long, it takes about 5 to 10 years of cycle, in the development, to gradually through the local mature technology to apply, try all the application scenarios.

One of the core concepts of Yishi Technology is to use the collaborative design of controller software and apply a core set of algorithms to different scenarios as much as possible.

The use of a set of algorithms on the engineering ability requirements are extremely high, in the rapid iteration of the algorithm, rapid evolution of the process, to consider the technology mature, engineering application, in order to quickly verify the technology, so that the technology mature.

At the same time, after the Robotaxi technology matures, it can be quickly applied to the scene to achieve a scene pattern that complements and iterates on each other.

2. Different technical routes of autonomous driving companies have their own strengths.

As we all know, in the field of autonomous driving, different companies have different technical routes. Some companies will launch L3-class models first, on the one hand, they can get revenue by selling cars, and on the other hand, they can collect more data by going on the road. Others, such as Tucson Future, have been developing L4-level autonomous driving technology from the start.

In response to this situation, Hao Jiannan, co-founder and chief architecture officer of Tucson Future, said that there are multiple exploration directions in the field of autonomous driving, and it is reasonable to gradually take the technical route and directly do L4 level automatic driving.

From a business perspective, the fundamental problem for freight transport to solve is to reduce operating costs.

L4 autonomous driving reduces costs by eliminating drivers, the largest component of current operations. The logic of the L2 or L3 autonomous driving route is to reduce the fatigue of the driver by applying assisted driving technology and thus reduce the investment of manpower.

L2 and L3 level automatic driving have great significance in terms of fuel consumption and fuel saving. Because driving by human power, it is difficult to achieve the same degree of fuel economy as autonomous driving controlled by the system. Through the automation of the system, fuel consumption can be controlled within the most economical range. Human drivers, on the other hand, have a hard time staying like this all the time.

From a technical point of view, L4 level has more data requirements, because L4 level to achieve a very high availability through system design, can not only rely on data accumulation, but also need some redundant design.

Dong Jian, co-founder and vice president of software algorithms, believes that the L2-level assisted driving system has a certain role in promoting the development of high-level automatic driving systems.

Dong Jian said that the ultimate goal of Hongjing Intelligent Driving is also an advanced automatic driving system such as L3 and L4. However, it first chose a gradual and sustainable path of autonomous driving technology.

The L2 automatic driving system has great significance for the L4 automatic driving system. First of all, there is no relatively complete regulation for L3 or above in China, that is to say, all the autonomous vehicles on the open road are actually L3 or below the automatic driving system. By choosing the route of technology iteration, companies can generate revenue at an early stage.

On the other hand, at present, more than 150,000 to 250,000 high-end models in China pay more attention to the automatic driving experience of L3 and L4, and car manufacturers will also do similar functional experiences. Like Tesla's hardware embedding and software OTA upgrade methods, it can provide a sustainable upgrade path.

In addition, we can accumulate a lot of mass production experience from L2 mass production projects, which are very important for L3 and L4 level autonomous driving systems.

Dong Jian said that the autonomous driving system is very complex, involving software, hardware and system integration and other aspects. Without the experience of mass production projects and the deep running-in with the main engine factory, it is more difficult to create an automotive product that can really be mass-produced.

3. Autonomous driving will mature after 2025.

Dai Zhen, vice president of Heduo Technology, believes that different users have different needs for autonomous vehicles, OEMs pay more attention to the interaction of automatic driving, users pay more attention to the interaction of the cockpit, and pay more attention to comfort, so the passenger cars mass-produced by autonomous driving companies should have both B-end and C-end users.

At present, this is also a bottleneck facing the commercialization of autonomous driving - to meet the dual demands of both B-end customers and C-end customer experience.

In addition, consumers' concerns about autonomous driving technology have also restricted the commercialization to some extent.

Dai Zhen believes that 2025 may be an important time node in the field of automatic driving, from the current point of view, before 2025 in some relatively specific parks, some closed, less crowded and low speed places, automatic driving has begun to be gradually applied.

Dai Zhen believes that for mobile travel on open roads, the spring of autonomous driving should be after 2025. The main reason is that the technical problems at the car end, including sensor and chip technology, are gradually moving toward mass production and commercialization. In addition, ordinary consumers are gradually increasing their acceptance of autonomous driving.

Finally, some supporting facilities, including infrastructure transformation and the improvement of laws and regulations, may take two to three years to mature. After 2025, when the above factors are truly implemented, automatic driving technology can really land quickly.