Baidu Taps 50 Partners And Details Its Autonomous Driving Project

摘要: On July 5th, Baidu announced the establishment of Apollo project, one of the largest partner ecosystems for an autonomous driving platforms around the world. What's the project about? What role will it play? What impact can it have on the Chinese or global automatic driving industry?

(Chinese Version)

On July 5th, Baidu held an AI Developer Conference, the biggest in its history. At the conference, Lu Qi, vice chairman of Baidu’s board of directors, president and COO of Baidu Group, revealed for the first time to the public Baidu’s grand vision in its Apollo ecosystem and announced the establishment of Apollo ecosystem project’s partnership alliance (if that name seems familiar, it’s because it’s actually named after the U.S. mission to the moon).

“Apollo will become a milestone for China’s auto industry. What it means for the Chinese auto industry is what Android means to the smartphone industry. However, its is even more open and potent,” Lu explained.

He was quite excited in his introduction to the project, which is known to be the most high-anticipated part of Baidu’s entire AI strategy.

As a matter of fact, Lu Qi announced the project as early as this April during the Shanghai Auto Expo, saying that Baidu would build an open, comprehensive and safe software platform for the auto industry and provide its partners an integrated hardware, software and service system in the self-driving sector.

At that time, Lu already pointed out that Baidu would do whatever it could to help players in self-driving sector avoid distracting efforts and put more efforts in developing an entire self-driving system of their own in a more effective and efficient manner.

Within only three months, Baidu has advanced its plan to an extraordinary degree thanks to Lu’s strong execution power and Baidu’s technological accumulation in the self-driving sector.

Over 50 partners of the Apollo project

Take a close look, and you will find that Apollo’s ecosystem is quite wide: in fact, over fifty partners are included in the first batch, covering almost all relevant parties along the unmanned driving industry chain, including manufacturers, universities and government departments. It is safe to say this is the most “luxurious” self-driving flagship all over the world.

Baidu’s partners include: over 15 OEMs (Chery, FAW and Chang’an, etc.), over 10 first-tier auto parts manufacturers (BOSCH Automobile, Continental Automobile, ZF Friedrichshafen, DESAY SV Automobile, etc.), over 12 key parts manufacturers (NDIDIA,Microsoft, ZTE, Velodyne, TomTom, Unigroup Spreadtrum & RDA, Forte, etc), three auto startups (Autonomous Stuff, DPXCN, etc.), three ride-sharing service providers (UCAR, Grab Taxi, etc.), five key universities (Tsinghua University, Shanghai Jiao Tong University, etc.) and six government departments and bureaus (Wuhu, Baoding, Beijing Yizhuang District, etc.).

It is reported that six out of ten top-tier OEMs have joined the alliance. BOSCH Automobile and Continental Automobile, two of the most well-known first-tier auto parts manufacturers,, have entered into strategic agreement with Baidu, as first-batch members of the Apollo ecosystem.

To everyone’s amazement, Baidu even issued Apollo’s strategic structure and specific business modules at the conference. This is also the first time Baidu revealed to the public its business project in a comprehensive and detailer manner. Therefore, how will Baidu’s alliance affect the industry?

Four layers of the Apollo project

Wang Jing’ao, senior director of Baidu and R&D manager of Apollo, introduced the four layers of the Apollo ecosystem: cloud service platform, open software platform, reference hardware platform and reference vehicle platform and claimed that Baidu would open up these four layers step by step:

The first layer is about cloud-based services, including high-precision mapping and navigation, autopilot simulation engines, data platforms, security services, OTA online upgrades, and DuerOS, which was described in detail at the conference.

The second layer is about software, including map engine, vehicle positioning, environmental awareness, path planning, end-to-end signal transmission, human-machine interaction, operating framework and real-time operating system that provide operating environment for all of these software.

The third layer is about hardware, including the computing unit (CPU/GPU/FPGA), positioning unit (GPS/IMU) and sensors for environmental perception (camera, radar, laser radar and other hardware).

The fourth floor is about vehicles. Up to this point, Baidu has established cooperative relationship with several well-known car makers, including Daimler, Ford, Chery, BAIC, Lincoln, GMW, etc.

Open pathway of the Apollo Project

As to the various capabilities and resources to be opened by the Apollo Project, Baidu has given a specific timetable.

In the Apollo Version 1.0 launched this July, Baidu’s close area tracking automatic driving capability is opened to the, as to cloud service, Baidu will open its intelligent vehicle system DuerOS; as to software platform, Baidu will open its self-positioning, control, HMI and many other technologies; at to hardware, Baidu will open its reference hardware, which is essential for developing automatic driving research and development; as to vehicles, Apollo provides a reference vehicle platform.

By this September, Baidu Apollo will open its fixed-lane automatic driving capability and part of its simulation engine data. By the end of this year, Apollo will open a series of new capabilities which will enable vehicles to complete autopilot tasks in simple urban traffic situations. In addition, Baidu will open up more data and data upload interfaces to the public.

TMTPost found outside the exhibition hall that: two modified Lincoln MKZ started and completed highly-difficult “Arabic number 8-like” spiral moves and reached the destination together. This is the first time two unmanned vehicles passed the simultaneous test in closed areas.
Close area tracking automatic driving capability

Close area tracking automatic driving capability

Modified Lincoln MKZ

Modified Lincoln MKZ

This is actually something developed by American cutting-edge automatic driving enterprise AutonomouStuff based on GPS and camera perception technology. According to Robert Hambrick, CEO of AutonomouStuff, it took only three days to build this set of automatic driving system with the help of Apollo 1.0’s open resource capacity.

Apollo 1.0 also provides global developers with a full range of “equipment”, including high-efficiency and easy-to-expand structure, complete and easy-to-use reference vehicles and hardware, one-on-install software modules, and complete off-line online development and debugging tools.

Gu Weihao, general manager of Baidu intelligent car business division, specifically introduced the three major capabilities at the conference, namely: simulation engine, high-precision mapping and positioning, security.

Simulation engine:  As one of the most important components of Apollo, simulation service accumulates a large number of actual traffic and unmanned scene data. Based on Baidu’s large-scale cloud computing capacity, it could simulate a virtual running capacity of millions of kilometers. Through open simulation services, Apollo’s partners can access unmanned scenes and quickly complete a series of tasks, including testing, verifying and modeling optimization.

High-precision mapping and positioning: Baidu is one of the pioneers that apply deep learning and AI technology to mapping data generation and one of the first few high-precision mapping and positioning service providers. Based on GPS, IMU, high-precision maps and a variety of sensor data, Baidu’s self-positioning system can provide centimeter-level integrated positioning solution.

Security: Apollo provides a complete security framework and safety system components in an isolated and reliable security system. Through the vehicle safety firewall, Apollo isolates the network within the outside of the vehicle, the network of the car body and the vehicle, which ensures the independence of network boundaries.

With the increasing number of partners inside Apollo ecosystem, shared data, resources and technology will multiply, which means that developers and enterprises will no longer need to do basic work, but instead put more labor and financial resources into more advanced R&D work, which corresponds to the vision mentioned by Lu Qi previously.

Various resources are gathering towards the Apollo project

Besides Apollo 1.0, Baidu will increasingly open Baidu’s driverless technology and capital to the outside.

Gu Weihao also introduced to developers and partners Baidu’s project in environmental perception, end to end, high-precision maps, simulation maps and other industries that can be empowered by the AI technology.

Apollo 1.0 opens access  its co-testing high-precision mapping technology. Through artificial intelligence, depth learning and other technologies, Baidu and international top-tier high-quality map service providers TomTom enhance the automation and intelligence of high-precision mapping data generation to meet the needs of automatic driving.

At the conference, Xu Yong, general manager of Baidu IoV Business Division, introduced the application of Baidu Intelligent Dialogue Artificial Intelligence System DuerOS in the automotive scenarios. According to him, Baidu will provide its partners with intelligent car solutions from various aspects, including product, technology and service, enabling enterprises to respond quickly to market demand and the change in automatic driving technology.

However,  government forces play an irreplaceable role in the perfection of self-driving facilities in the era of self-driving cars. On July 5th, Baidu has signed a strategic cooperation agreement with Baoding city government, agreeing to jointly build an intelligent transportation city. Besides Baoding, five other regional government, including Wuhu, Baoding, Beijing Yizhuang District, Chongqing Liangjiang New District, Shanghai Automotive City, are also included in the first batch of Apollo ecosystem’s alliance.

The only open technology platform till now

Currently, the self-driving industry is facing various problems, including industrial structure change, technology itself and supporting laws and regulations, etc. As an open technology platform, Apollo is a starting point  of prosperity built together by developers and partners. The ultimate goal is to promote the development of intelligent driving, and provide high-efficiency, secure solutions for each partner.

According to Baidu, Baidu will also promote the industry-chain development of Apollo ecosystem’s partners  in the future, gather relevant players in every phase, open up the entire industry chain and explore updated cooperation mode.

While the automobile industry chain itself is very complex, related technologies and phases to achieve automatic driving are also very complex. At present, many alliances have already been established among R&D parties, suppliers and tech companies, such as the alliance built by Intel, mobileye , BMW and Delphi. However, all relevant players are basically still buried in themselves, and there is still no real open technology  platform before Apollo project.

From this aspect, the emergence of the Baidu Apollo project to breaks the deadlock. Perhaps, this project is still insufficient in the early stage. However, it already attracted a fair number of partners to make it happen. In fact, such cooperation could grow into a vital way to accelerate the popularization and commercialization of self-driving technology.

However, it is worth noting that among the 50 partners of Apollo 1.0, many of them are actually competitors. Perhaps, whether Baidu can coordinate their interests, allow car makers, suppliers and tech companies to work together and achieve prosperity within the framework of the Apollo project, will determine whether it could grow into a preferred tech platform in the driverless technology sector.

……………………………  

(Like our Facebook page and follow us now on Twitter @tmtpostenglish, on Medium @TMTpost, on Instagram @tmtpost_english and on Apple News@TMTpost)   

[The article is published and edited with authorization from the author @Ti Geek please note source and hyperlink when reproduce.]    

Translated by Levin Feng (Senior Translator at PAGE TO PAGE), working for TMTpost. 

更多精彩内容,关注钛媒体微信号(ID:taimeiti),或者下载钛媒体App

本文系钛媒体原创,未经授权不得使用,如需获取转载授权,请点击这里
分享到:

第一时间获取TMT行业新鲜资讯和深度商业分析,请在微信公众账号中搜索「钛媒体」或者「taimeiti」,或用手机扫描左方二维码,即可获得钛媒体每日精华内容推送和最优搜索体验,并参与编辑活动。

tmtpost
tmtpost

Our official account in English/English Version of TMTpost.com

评论(0

Oh! no

您是否确认要删除该条评论吗?

分享到微信朋友圈