From Mocha to Tank 500, how did the Hpliot go from 1.0 to 2.0 in half a year?

Recently, Great Wall Motors' tank 500 thousand calls began to come out, and with its listing, the outside world's view of luxury SUVs has changed. Taking into account both off-road and urban travel, tank 500's "poetry and distance" and "quality of life" have found a reconciliation in intelligence.

It is reported that the tank 500 is the first model of the HPilot 2.0, and its auxiliary driving has added a new function of berthing. The Tank 500, which emphasizes driving experience and off-road performance, adds the latest configuration for smart driving, a change worth playing with.

HPilot 1.0 iteration to HPilot 2.0 took only half a year

The HPilot 1.0 was just mass-produced last year, and in less than half a year, HPilot 2.0 was on the train, which can be described as rapid progress.

What role can HPilot 2.0 play on the Tank 500? According to reports, the HPilot assisted driving system is the first domestic leading functional design, comprehensive scene coverage, and large-scale mass production of assisted driving solutions.

Since May last year, the HPilot 1.0 has been successively installed in five Great Wall models of Weipai Mocha, Tank 300 City Edition, Weipai Machiduo DHT, Weipai Latte DHT, and Haval Divine Beast, and in October 2021, it became the third manufacturer in the world to officially launch a full-stack self-developed intelligent pilot assistance system NOH.

Wang Tianpei, R&D director of Zhixing, introduced that compared with HPilot 1.0, HPilot 2.0 not only has a stronger computing power, but also adds the "parking integration" driving and parking intelligent scene integration solution.

Its parking solution includes fully automatic and remote parking, as well as a follow-up function. Through the optimization of forward-looking perception and other upgrading of the perceptual level, it is possible to see a God perspective composition when parking. When automatic parking, the working conditions are less than 30km/h, the size of the side parking space is greater than the length of the body 0.8m, the vertical parking space size is greater than the width of the body 1m, and then press the button, the system will calculate the best parking trajectory, and provide the driver with a parking prompt through the display or voice.

The memory route length of the memory parking function is planned to be greater than 1km, and there are more than 50 routes, and cross-floor parking is still being planned.

In the daily commuting and long-distance travel scenarios, the system can provide five highlight functions: intelligent cruise, autonomous lane change, intelligent cornering, intelligent dodge, and automatic emergency braking. It has more than 30 assisted driving functions, which can effectively assist users to automatically complete frequent driving operations, making driving easier and safer.

The Tank 500 is positioned luxuriously, so the advanced intelligent driving features are not at all contrary to it. According to Wang Tianpei, the HPilot 2.0 will be listed on other models.

From the mass production of HPilot 1.0 in the middle of last year to the mass production of HPilot 2.0 in the first quarter of this year, the reason why HPilot can achieve such an upgrade speed is due to talent. Under the outlet, autonomous driving talents are in short supply, especially the importance of self-developed car companies to autonomous driving talents, and they do not hesitate to attract people at great expense.

In this regard, there is a certain accumulation of Millima Wisdom, Wang Tianpei revealed that there are more projects under research at present, and now the R & D personnel at Milliper account for nearly 90%, and there are nearly 100 R & D personnel responsible for HPilot upgrade projects.

He believes that the attraction of Zhixing to talents comes from the company's mass production ability and talent aggregation effect. "Having a real landing project within the enterprise, because just doing some research and really doing a project that may take off the hat is completely different, which is very helpful for everyone's growth." At the same time, the company has a lot of high-end cattle people, and it can attract some potential young people to do it together. ”

It is understood that in the process of upgrading HPilot from 1.0 to 2.0, on the one hand, it is necessary to coordinate the collocation with the organization's human resources, on the other hand, to develop on the chip with greater computing power, it is necessary to re-establish the development capabilities of the underlying layer, and then at the same time to make the underlying more stable, which also takes a certain amount of time.

Intelligent driving systems have become the focus of competition for new cars

Tank No. 500 in the domestic fuel SUV, its auxiliary driving function is the leading. In fact, among the mass-produced models, there are not many models that really have intelligent auxiliary driving in the driving domain and parking domain, most of which are represented by new power brands, such as Weima that uses Baidu AVP, Xiaopeng who develops xpilot, etc.

The Great Wall's fuel model is equipped with an intelligent driving system that integrates the parking, and it is the first time to board the car in the latest version, which also indicates that the status of intelligence is increasing.

In the process of chasing intelligence, different car companies have different solutions. Tesla, Wei Xiaoli and other efforts to maintain the self-research route, second-tier electric vehicle brands to baidu, Huawei and other technology companies technology platform, Great Wall incubated independent companies specializing in self-research.

Today, it seems that HPilot has benefited from the endorsement of the Great Wall, and the plan has been quickly landed. At this stage, intelligent driving has reached the stage of mass production of PK.

There is data for mass production, and the more data accumulated, the more it can feed back technological innovation. Wang Tianpei said that the current automatic driving technology mainly relies on three changes: the first is the computing platform, but there are not many choices, and the main providers include NVIDIA, Qualcomm, and some new domestic forces. The difference in this part will not be particularly large.

The second point is the algorithm, the establishment of the algorithm requires high-level talents, such as through the industry bull to quickly establish, to make up and accumulate.

In addition, the most important point is data, data can not be established overnight, it must have a source of data, and relatively good data processing process, and then can quickly accumulate data, and finally serve the algorithm.

"For example, if you only use a few test cars to run, it is certainly better to install the system on the model with sales, as long as the main engine factory sells the car, there will be a data running on the road." Wang Tianpei said.

As of April 2022, the driving range of the user assistance of Miller Has exceeded 7 million kilometers, and it is expected that in the next two years, the total number of passenger cars equipped with HPilot assisted driving system will exceed 1 million units, if this goal is achieved, it means that Millais Wisdom will become the industry's shortest time to enter the "Million Mass Production Club" of autonomous driving companies.

Compared with other self-driving companies, The Great Wall-backed Zhixing has been able to catch the mass production express.

Autonomous driving startups are actively seeking to cooperate with OEMs to get a larger scale of landing, but Zhixing has taken shortcuts. Wang Tianpei said: "We save the process of spending a long time to acquire customers, and there may be some startups that are doing well, but they still have to spend more time or more cost to find some mass-produced customers, and then implement their own systems." ”

Che Yun summary

Today's intelligent driving function has increasingly become the highlight of the upgrade of new cars, some time ago, Li Xiang suggested "making AEB the standard of the whole vehicle" and won the approval of Miao Wei. Of course, the scope of AEB is very wide, but the trend of automotive intelligence has become unstoppable.

Although we are still a long way from becoming fully autonomous, assisted driving is advancing by leaps and bounds. At the policy level, Mercedes-Benz has already implemented L3 road access permits on German roads, which means that in the event of a system accident, the party responsible will be the manufacturer. But Wang Tianpei believes that in China, in the more complex environment of road conditions, the development of intelligent driving is more concerned with accumulating data and constantly breaking through the corner case, and is not so concerned about defining the level.

With the goal of tackling key scenarios, intelligent driving is sweeping towards large-scale mass production.