EasyCar News? "For the many problems that exist in autonomous driving, we should think about their nature with the first nature principle, from which it is possible to make different products." On July 8, the CEO of the milli-millimeter Zhixing Gu Weihao was invited to participate in the 2022 14th China Automotive Blue Book Forum, and the "user needs what kind of automatic driving" to start the exchange of discussion. Gu Weihao said, the continuous progress of the product ability, there is only one way, is the drive of big data, sitting in the office by a few people's vision, can not make a good product.

This year's forum, themed "Refresh", featured more than 60 keynote speeches and 12 panel discussions. The conference invited hundreds of the most thoughtful leaders and elites in China's automotive industry, including Dongfeng Motor, SAIC, BAIC, FAW Group, Great Wall Motor, Chery Automobile, Daimler, Azure, Xiaopeng, Baidu, Huawei, Horizon, etc., to contribute to the industry's development of the annual feast of ideas and brainstorming storms.

Users are most concerned about the product experience

Different from the previous Autonomous Driving Forums, which were focused on the technology itself, this year's China Automotive Blue Book Forum focused on the question of "what kind of autonomous driving do users need?

The year 2021 has been hailed as the first year of commercialization of autonomous driving, with major car companies and head autonomous driving companies commercializing their high-speed domain advanced assisted driving products; in 2022, with the competition for autonomous driving technology moving from high-speed to urban domains, the research and development of advanced assisted driving technology for urban domains has also entered the countdown.

Despite the industry's optimism about the development of autonomous driving, there is an obvious lack of user perception. Specifically, the L3 level of autonomous driving has always remained in the conceptual stage, while around the L2 level of assisted driving such as L2 +, L2.5, L2.99 and other marketing packages are endless, is this out of the need for corporate publicity, or is it really able to bring qualitative changes in the user's travel experience?

Gu Weihao expressed his views on this. He believes that the main difference between L2 assisted driving and L3 autonomous driving lies in the division of responsibility; for users, the latter cares more about the product experience. Gu Weihao talked on the scene, "I, as a user of millimeters Zhixing assisted driving products, drive Wei brand Moca and Tank 300 to and from work every day, and what I care most about in this process is the product experience and improvement."

"During the driving process, my task is to supervise my car driving by itself, and I hope that I intervene less and less so that I can know what scenarios it is suitable for and what scenarios it is not suitable for, and that there are some scenarios where I can just relax a little bit, and there are some scenarios where I can be a little bit more urgent." Weihao Gu said assisted driving products are shifting the role of the driver from "manipulator" to "supervisor".

How do users perceive the differentiated experience brought by different levels of autonomous driving capabilities? Gu Weihao mentioned, the user does not care what you provide is L several, L several have applicable scenarios, the user used familiar naturally know which section of the road can drive, slowly so drive. With the continuous upgrading of the vehicle, the user's own feelings can also be upgraded and increased. Now, the user's high-frequency auxiliary driving of the core use of the scene is to follow the scene, the intersection scene and change lanes scene.

And the key to breaking through to autonomous driving specifically lies in data. Thanks to the Great Wall Motor's advantage of large-scale mass production, Gu Weihao said, Milliwheel has obtained the key to winning the rapid iteration of automatic driving, and thus launched the data intelligence system MANA, to accumulate data at high efficiency and low cost as the core, driving the continuous iteration of automatic driving technology.

Currently, MANA's assisted driving products have been mass-produced on a large scale in six Great Wall Motor models, and users' cumulative assisted driving mileage has exceeded 10 million kilometers, which will grow exponentially along with the ability to scale mass production. "Autonomous driving requires a larger vehicle scale and data scale, in addition to a stronger ability to convert data into knowledge, so that autonomous driving can quickly realize the iteration from quantitative to qualitative change, thus growing to L3, or even L4."

Millimetre CityNOHMass production sprint is imminent First nature of the principle of starting millimeters to choose the route of heavy perception

A successful product can not only focus on the technical level, but also consider the user's aspirations, based on the depth of the first principle of thinking, Gu Weihao on the development direction of the autonomous driving Based on his deep thinking on the first nature principle, Gu put forward his own viewpoint on the development direction of automatic driving.

In Gu Weihao's view, urban commuting has become the most important scenario for automobile use, and users' demands for advanced assisted driving products in the urban area have become clearer, i.e., safety, simplicity, and ease of use.

What should an urban advanced assisted driving product be like that meets users' demands? Before this, many automatic driving companies have put the core of the competition on the perception hardware and high precision maps, but Gu Weihao put forward the development trend of the automatic driving industry, expressed his own views, "the reason why the high precision maps and LiDAR first appeared is because at that time, the camera perception ability of the automatic driving technology was relatively weak, and the performance of the arithmetic power is also not enough to support the Perception algorithm; today not only the camera pixels have been greatly improved, big computing power chips also began to emerge, perception ability has been greatly improved, and even the cost of expensive LiDAR cost gradually began to dip."

Thus, Milliwheel puts forward the "heavy perception" technology program. Its data intelligence system MANA as the core, through the massive data feeding, so that MANA learning time of more than 240,000 hours, the virtual driving experience is equivalent to the human driver driving 20,000 years, which allows milliMax city NOH feel and master to deal with all kinds of complex urban scenarios driving know-how, so as to realize from the "novice" to the "old driver". This will enable NOH to realize the transformation from a "novice" to an "experienced driver", allowing drivers and passengers to enjoy safe, comfortable and efficient travel.

Excellent user experience is not only limited to safety, comfort and efficiency, in fact, there are still a lot of details in the process of using self-driving products that deserve attention and improvement from self-driving companies. In this regard, Gu Weihao pointed out a typical example, "When a vehicle with high-speed NOH is turned on and passes a large truck, the driver will often instinctively take over the vehicle because of the fear of the big car, even though the vehicle will not collide with the big car at this time."

In this regard, Gu Weihao introduced the function of milli NOH "Smart Dodge", "If there is a large vehicle on the right side, we will set the vehicle to shift to the left in the left lane by 10-15 centimeters. At the same time, milli NOH will be labeled in the interactive interface, informing the user of the vehicle's evasive action, giving the user psychological expectations." Gu Weihao mentions, "This feature would initially use analog artificial voice to remind users, but when a large number of large vehicles are encountered on the road, the frequent prompting voice would bring more trouble to users. Then we changed this interaction to a low, short voice, which makes it easier for users to accept than the previous voice reminders." This case shows that by paying attention to and adjusting the details, MilliMax Zhixing has gradually made its assisted driving products more mature and is gradually differentiating its products.

Gu Weihao's clear and unique viewpoints also triggered discussions and heated debates among many guests at the scene. This session of the China Automotive Blue Book Forum this year to "refresh" as the theme, behind the high-speed development of the industry, hand in hand to break the boundaries of innovation, through the cycle.