Home > Technology peripherals > AI > Gu Weihao, CEO of HaoMo Zhixing: Six major challenges and new upgrades faced by MANA in urban scenarios

Gu Weihao, CEO of HaoMo Zhixing: Six major challenges and new upgrades faced by MANA in urban scenarios

WBOY
Release: 2023-04-14 11:58:02
forward
1170 people have browsed it

"The era of data-driven autonomous driving 3.0 has arrived. Only when the four technical conditions of perception, cognition, mode, and data are established in parallel can we truly enter the new era of autonomous driving. Everything we do is to enable Create data channels and computing centers so that data can be acquired more efficiently and transformed into knowledge. BiMo is sprinting into the autonomous driving 3.0 era with all its strength!" At the 6th HAOMO AI DAY held on September 13, BiMo Dr. Gu Weihao, CEO of Zhixing, delivered a speech titled "The 3.0 Era of Autonomous Driving and Autonomous Driving". For the first time in the industry, he proposed the industry judgment that "autonomous driving has entered the data-driven 3.0 era". He also introduced the MANA data intelligence system Based on the exploration and layout of cutting-edge AI technology, major functional upgrades in autonomous driving urban scenarios.

Gu Weihao, CEO of HaoMo Zhixing: Six major challenges and new upgrades faced by MANA in urban scenarios

(Dr. Gu Weihao, CEO of HaoMo Zhixing, delivered a keynote speech on "HaoMo and the 3.0 Era of Autonomous Driving")

Embrace the Attention big model, Continue to maintain technological leadership and sprint towards the autonomous driving 3.0 era with all our strength

What is the autonomous driving 3.0 era? What are the driving factors? What stage is Hao Mo currently in? Gu Weihao shared the above issues in his speech.

Gu Weihao, CEO of HaoMo Zhixing: Six major challenges and new upgrades faced by MANA in urban scenarios

(Gu Weihao’s live speech)

Gu Weihao said that the Attention large model is a new trend in the current development of AI and the opportunities it brings and challenges, becoming one of the important driving factors in the autonomous driving 3.0 era. The biggest feature of Attention is that it has a simple structure and can infinitely stack basic units to obtain a model with a huge number of parameters. With the increase in the number of parameters and the improvement of training methods, the effect of large models has surpassed the average human level in many NLP tasks. However, Attention's large model also faces a big challenge, that is, because its demand for computing power far exceeds Moore's Law, which results in very high training costs for large models and very difficult to implement on terminal devices.

Gu Weihao, CEO of HaoMo Zhixing: Six major challenges and new upgrades faced by MANA in urban scenarios

(Gu Weihao said that the Attention large model is a new trend in the current AI development)

The opportunities and challenges brought by the Attention large model are driving Technological changes in the autonomous driving industry. "Haimo is reducing the cost of autonomous driving through low-carbon supercomputing, implementing large models on the vehicle side by improving the design of car-side models and chips, and making large models more effective through data organization." Gu Weihao said, At the data level, based on the Attention large model, autonomous driving requires large-scale and diverse training data. Only passenger car assisted driving based on large-scale real human driving data has the ability to accumulate sufficient scale and sufficient diversity of data. "We have reason to believe that assisted driving is the only way to autonomous driving. Because only assisted driving can collect data of sufficient scale and variety." It is reported that after nearly three years of development, Weimo is now China's largest volume As the number one autonomous driving company in the industry, the current user-assisted driving mileage has reached nearly 17 million kilometers, and the data scale continues to increase rapidly.

(Gu Weihao said that assisted driving is the only way to autonomous driving)

At the low-carbon supercomputing level, Wei Mo officially announced China’s autonomous driving technology company at this AI DAY The first supercomputing center. Gu Weihao said: "How to improve training efficiency, reduce training costs, and achieve low-carbon computing is a key threshold for autonomous driving to enter thousands of households." The goal of the Haimo Supercomputing Center is to meet the needs of large models with hundreds of billions of parameters and training data scale With 1 million clips, the overall training cost is reduced by 200 times.

Gu Weihao, CEO of HaoMo Zhixing: Six major challenges and new upgrades faced by MANA in urban scenarios

(China Autonomous Driving Company’s first supercomputing center - Haimo Supercomputing Center unveiled)

At the algorithm model level, Gu Weihao introduced , Haimou launched research and implementation attempts on the transformer large model as early as June 2021. It is based on the successful practices of the past year and more in the transformation and upgrading of training platforms, preparations for switching data specifications and annotation methods, and exploration of model details for specific tasks of perception and cognition, etc., that we have laid the foundation for the current situation of Hao Mo in urban navigation assisted driving scenarios. Rapid development has laid a solid foundation.

Gu Weihao, CEO of HaoMo Zhixing: Six major challenges and new upgrades faced by MANA in urban scenarios

(Hao Mo’s new technology practice path)

"The era of data-driven autonomous driving 3.0 has arrived." Gu Weihao believes that the development of autonomous driving in the past ten years can be divided into three eras: hardware-driven, software-driven, and data-driven. The data-driven era is a completely different era. Large models and massive data are a "double sword", and the data turns on the self-training mode; in perception technology, multi-modal sensors jointly output results; in cognitive technology, interpretable scenarios are used Driving common sense is the main thing; autonomous driving mileage is driven by hardware and software. The mileage in the era of millions of kilometers and tens of millions of kilometers has soared to more than 100 million kilometers. With data drive as the core, only when the above four technical conditions are established in parallel can we truly enter the era of autonomous driving 3.0.

"Hai Mo has been preparing for the autonomous driving 3.0 era. In terms of perception, cognition, and model construction, it is all constructed in a data-driven manner. Everything we do is to be able to do Data channels and computing centers can be used to obtain data more efficiently and transform data into knowledge." Currently, Tesla has led the world in entering the era of autonomous driving 3.0, and Feimo is most likely to become the first Chinese company to enter the era. A company in the autonomous driving 3.0 era.

“We are passionate about innovation, open to new ideas, new methods, and new technologies, and pay special attention to technical routes that can form a positive cycle with the growth of data scale. This is also the first priority in making technical strategic decisions. Principle: The technical route that can quickly transform the advantage of data scale into the advantage of capabilities is a good route." Gu Weihao said that when it comes to the exploration and implementation of cutting-edge technologies, Wei Mo will always maintain the most extreme, keenest and most open mind, and strive to Provide users with better product experience and promote the development and progress of the industry.

MANA's six major milestone upgrades lead the industry in opening the way for urban NOH to "enter the city"

The ultimate pursuit of leading technology not only allows Haimo to always go At the forefront of industry innovation, it also takes the lead in technology exploration and implementation in urban assisted driving scenarios where autonomous driving companies are collectively competing.

The urban navigation assisted driving scenario is the core breakthrough point of the current autonomous driving function, and it is also a battleground for military strategists. However, from a high-speed scene with single roads and traffic conditions to an urban scene with many traffic participants and extremely complex road and traffic conditions, the technical difficulty faced by the autonomous driving system can be said to have increased exponentially. Huge challenges have also hindered the pace of many autonomous driving manufacturers' "entering the city", and they can only continue to fight fiercely for technological breakthroughs. As China's number one mass-produced autonomous driving manufacturer, Feimo has set its flag to win the "battle for assisted driving urban scenarios" as early as the end of 2021, and has taken the lead in launching a journey of technological exploration in the field of urban assisted driving. Now Feimo Data Intelligence System MANA is ushering in a number of milestone upgrade iterations.

Gu Weihao said that urban roads mainly present “four types of scene problems and six major technical challenges.” The scene problems mainly include "urban road maintenance", "dense large vehicles", "narrow lane changing space" and "diverse urban environment". To solve the above scenario problems, the technical level faces six major challenges: how to convert data scale into model effects more efficiently, how to make data play a greater value, how to use re-sensing technology to solve the problem of real space understanding, how to use the human world's Interactive interfaces, how to make simulations more realistic, and how to make autonomous driving systems move more like humans.

In order to cope with the above challenges, MANA’s sensory intelligence, cognitive intelligence and other aspects have been updated and upgraded.

First of all, MANA uses a self-supervised learning method using unlabeled data of large-scale mass production vehicles to create model effects. Compared with training with only a small number of labeled samples, the training effect is improved by more than 3 times, which allows the advantage of milliseconds data. Efficiently transform into model effects to better adapt to various perception task requirements of autonomous driving.

Secondly, MANA’s perception capabilities have been improved, so that massive data will no longer be treated differently. Faced with the problem of "data efficiency" under huge data scale, MANA built an incremental learning and training platform, extracting part of the existing data and adding new data to form a hybrid data set. During training, the output of the new model and the old model are required to be as consistent as possible, and the fit to the new data is as good as possible. Compared with conventional methods, the overall computing power is saved by 80%, and the response speed is increased by 6 times.

Third, MANA has stronger perception ability. By using the time-series transformer model to perform virtual real-time mapping in the BEV space, the output of perceived lane lines is more accurate and stable, allowing autonomous driving in urban navigation to bid farewell to high-precision maps.

Fourth, MANA’s perception capability is more accurate, so there are no vehicle signals that cannot be recognized in China. By upgrading the on-board perception system, MANA specifically identifies the status of brake lights and turn signals, allowing drivers to be safer and more comfortable in scenarios such as sudden braking or emergency cuts of the vehicle in front.

Fifth, MANA’s cognitive ability has also evolved again. Facing the intersection, the most complex scene in the city, MANA introduced high-value real traffic flow scenes into the simulation system. It cooperated with Zhejiang Deqing and Alibaba Cloud to introduce the intersection, the most complex scene in the city, into the simulation engine to build an autonomous driving scene library. , through the real simulation verification of autonomous driving, the timeliness is higher and the microscopic traffic flow is more realistic, effectively solving the "old difficulty" problem of passing through urban intersections.

Finally, MANA cognitive intelligence has ushered in a new stage. Through in-depth understanding of a large number of human driving across the country, learning common sense and personification of actions, the driver's assisted driving decision-making is more like actual human driving behavior. It can choose the optimal route based on the actual situation to ensure safety, and the body feels more like an experienced driver.

The re-evolution of MANA has eliminated the biggest obstacle on the way to "entering the city" for Haomo City NOH. “Haimo City NOH is a navigation-assisted driving system that better understands China’s urban road conditions.” Gu Weihao said that Haimo City NOH adopts the technical route of “emphasis on perception, light on maps, and large computing power”. With the help of MANA, it has intelligent recognition The five highlight functions are traffic lights, intelligent left and right turns, intelligent lane changing, intelligent obstacle avoidance - static, and intelligent obstacle avoidance - dynamic. In addition, the "smart traffic flow processing" function will also be officially released.

At the end of the speech, Gu Weihao expressed the firm confidence and enthusiasm of the people in the future of autonomous driving. "More than 1,000 days ago, we were born at the end of the world and witnessed the fastest 1,000 days of autonomous driving in China. We are delighted with the achievements we have made. But 1,000 days is only the beginning of the battle. Let autonomous driving truly fly into the ordinary world. The homes of ordinary people are our stars and the sea. On the road to self-driving, everyone will continue to struggle and fight!"

The above is the detailed content of Gu Weihao, CEO of HaoMo Zhixing: Six major challenges and new upgrades faced by MANA in urban scenarios. For more information, please follow other related articles on the PHP Chinese website!

Related labels:
source:51cto.com
Statement of this Website
The content of this article is voluntarily contributed by netizens, and the copyright belongs to the original author. This site does not assume corresponding legal responsibility. If you find any content suspected of plagiarism or infringement, please contact admin@php.cn
Popular Tutorials
More>
Latest Downloads
More>
Web Effects
Website Source Code
Website Materials
Front End Template