Eight years ago, Alibaba JD.com's online shopping was not so developed, Tencent's WeChat was not yet popular, and Huawei mobile phones did not need to be snapped up. Baidu, which started with a search engine, has opened an industrial radar that moves with the wind and laid out automatic driving.
In other words, Baidu is the earliest Internet technology company in China to get involved in autonomous driving. A few days ago, Robin Li said at the Baidu AI Developer Conference that Baidu has become the world's largest provider of autonomous driving travel services. In the third quarter of this year, Baidu's newly upgraded self-driving travel service platform "Radish Run" provided 115,000 services. At the same time, Robin Li announced a small goal of turnip run - to expand the service to 65 cities by 2025 and 100 cities by 2030.
Even if the title of "the world's largest self-driving travel service provider" is aside, Baidu's answers in the field of autonomous driving in recent years are remarkable. However, in the context of the acceleration of technology giants to seize the beach, whether it is Baidu or other players, I am afraid that they need to speed up their pace to go further.
Eight years, the first shot of commercialization was fired
On November 25, Beijing officially opened the first pilot project for the commercialization of autonomous driving services in China, and Baidu became one of the first enterprises to be allowed to carry out commercial pilot services.
In fact, as early as May this year, Baidu Apollo opened its driverless service to the public in Beijing's Shougang Park. Even if there is a certain subsidy for paid services, Baidu can be said to have fired the first shot in the domestic Robotaxi commercial paid operation.
Even, Robin Li also gave Robotaxi's profit schedule - by 2025, Robotaxi's service cost will be lower than that of manual driving, and it is believed that it will achieve profitability with the expansion of Baidu's business scale. It's worth noting that this is also the first time that a self-driving player has proposed a Robotaxi profit plan.
In fact, whether it is the first crab to eat in the field of autonomous driving travel commercialization, or the courage to first propose the Robotaxi profit plan, this comes from the "ten years of work under the stage".
Wu Qiong, a manager of the Beijing Intelligent Vehicle Industry Innovation Center, said in an interview with industry media that in recent years, China has actively promoted autonomous driving road testing, so as to achieve experience accumulation and technology iteration, laying the foundation for the commercial operation of automatic driving.
Image source: Baidu Apollo
According to the third quarter financial report released by Baidu, Apollo L4 has accumulated more than 16 million kilometers of autonomous driving test mileage, an increase of 189% year-on-year, and in terms of test qualifications, Baidu Apollo has obtained 411 automatic driving test licenses.
More importantly, Baidu's accumulation in the field of automatic driving dates back to eight years ago. In 2013, Baidu established the deep learning laboratory IDL, and officially invested in the following year to launch the unmanned vehicle project.
As far as the background of the industry at that time was concerned, autonomous driving ushered in a wave of enthusiasm, and perhaps to some extent, it can be said that the times gave Baidu the opportunity to join the new track. On the one hand, traditional giants including General Motors, Ford and BMW are all developing self-driving technology. On the other hand, seeing that technology companies such as Google have made certain achievements in the field of automatic driving, Tesla, Apple, etc. have followed up to seize the autopilot market.
In 2015, Baidu Autonomous Driving Division was established. At the end of that year, Baidu's driverless car debuted on the road in China - realizing "full autonomous driving" under the mixed road conditions of cities, ring roads and highways, with a maximum speed of 100 km / h.
When the time came to 2016, Robin Li announced that Baidu would transform into an AI company, and said that it would build a complete AI open ecosystem in the future. As we later learned, autonomous driving is an important strategy for this open ecosystem. This year is an important inflection point for Baidu's automatic driving.
Pulling its sights back to 2017, Baidu announced through internal emails that it would merge its autonomous driving division, intelligent car division, and internet of vehicle business into an intelligent driving business group (IDG). A month later, at the Shanghai Auto Show, Lu Qi, then president of Baidu Group and COO, announced the Apollo plan to provide an open, complete and safe software platform to partners in the automotive industry and intelligent driving field, helping them to quickly build a complete intelligent driving system of their own by combining vehicles and hardware systems.
At that time, this not only represented more resource support, but also represented Baidu's automatic driving from relatively closed to open, attracting peers to build an ecosystem. At present, Baidu Apollo Autonomous Driving Open Platform has completed 11 iterations, with more than 80,000 developers in 135 countries, more than 210 partners, and an overall open source code volume of 700,000 lines, which has grown into the world's most active autonomous driving open platform. At the Baidu AI Developer Conference Autonomous Driving Sub-forum yesterday (December 29), Baidu officially released the Apollo 7.0 version.
Turning the time back to 2018, when the automatic driving track entered a small peak, Baidu also invested heavily in segments such as terminal logistics, lidar, Internet of Vehicles, and millimeter wave radar. For example, Baidu Autonomous Driving Division and Suning Logistics jointly released a cooperation plan. The two companies announced that they will jointly implement logistics autonomous driving technology to achieve the popularization of autonomous driving technology for terminal distribution and the large-scale mass production of unmanned delivery vehicles in 2020.
It was also in this year that the Uber driverless car died in an accident, becoming the world's first truly "self-driving car homicide." Subsequently, the domestic market also gradually realized that the mass production of automatic driving was not as expected, and the development of the industry entered a period of stability. Therefore, compared with the previous drastic moves, Baidu's investment in the field of autonomous driving in 2019 tends to be calm.
All-round attack, multi-scene layout
However, after this year's official announcement of the next car, Baidu in the field of automatic driving on the one hand to strengthen the layout of the industrial chain in an all-round way, on the other hand, to accelerate the multi-scenario landing process.
As far as the lidar segment is concerned, in August this year, Baidu's intelligent car manufacturing company Jidu Automobile and the domestic laser company Hesai Technology officially signed a contract, and the two sides will give full play to their respective technical and resource advantages to create a safer and more efficient intelligent driving system through in-depth cooperation.
Image credit: Velodyne
In fact, as early as 2016, when the high cost of lidar deterred competitors in the industry, Baidu had already invested in Velodyne, an established player in the field of lidar, and later also used lidar on Apollo driverless cars. In 2017, Baidu Apollo has jointly released the autonomous driving developer suite with Hesai, and the latter received 250 million yuan of financing from Baidu in 2018.
In terms of the pure visual path of automatic driving, Baidu can also be regarded as a senior player. During the CVPR2019 conference, Baidu Apollo unveiled the only self-driving pure visual city road closed-loop solution in China - Baidu Apollo Lite. At that time, Apollo Lite was the only urban autonomous driving pure visual city road closed-loop solution in China, which could support the parallel processing of 10 cameras and 200 frames per second data volume, and the maximum frame loss rate of a single visual link could be controlled below 5 ‰, realizing all-round 360 ° real-time environmental perception, and the stable detection of forward obstacles had a visual distance of 240 meters. In addition, at the 2021 Shanghai Auto Show, Apollo made its debut and focused on upgrading the LiDAR-free Pure Vision Autonomous Driving Solution (AVP+ANP).
In October this year, Baidu has twice invested in Shanghai Geometry Partners Intelligent Driving Co., Ltd. Geometry Partners is an integrated system provider of autonomous driving hardware and software systems focused on 4D millimeter wave imaging radar as the main sensing, supplemented by the fusion of visible light and infrared imaging multi-sensors.
In the past few years, few companies in the world have made breakthroughs in the development of 4D millimeter wave imaging radars, so the technology has not been popularized. However, when the time enters 2021, 4D millimeter wave imaging radar has suddenly become one of the focuses of the industry, especially under the role of Huawei and SAIC. Since the official announcement of the car, Baidu has invested in geometric partners twice, which to some extent reflects its emphasis on 4D millimeter wave imaging radar.
In the field of chips, Baidu has also been taking the path of self-development. At the Baidu World 2021 Conference, Baidu announced the mass production of Kunlun 2nd generation AI chips. Robin Li said that in areas that require large computing power, such as automatic driving, Kunlun chips will be useful.
At present, the global automobile industry is facing the dilemma of "lack of core", and building a complete vehicle-level ecological chain has become an urgent task for domestic chip companies. Incubating more chip suppliers and ensuring the chip supply of future smart cars has become the layout idea and investment opportunity of relevant enterprises and institutions. Then, Baidu naturally must also seize the opportunity.
In addition to strengthening the layout of the industrial chain, since the beginning of this year, Baidu Apollo and a number of ecological partners have jointly created "automotive robots" that run through the fields of passenger transportation, freight, and life services, and comprehensively land subdivided scenarios such as passenger cars, buses, trunk logistics, warehousing and distribution, and mine port operations.
In October, Baidu Apollo and Huaneng Yimin Coal and Power Company jointly announced the results of the phased cooperation in the field of unmanned mining cards. This marks that on the basis of the research and development results of 8 years of autonomous driving technology, Baidu Apollo has added another commercial landing scenario.
Also this year, DeepWay, a new force in truck manufacturing, launched the first all-forward design and development of intelligent new energy heavy trucks - Xingtu 1 generation, and DeepWay is a technology company jointly built by Baidu and Shiqiao.
Earlier this year, at the 8th International Intelligent Connected Vehicle Technology Annual Conference, the Beijing High-level Autonomous Driving Demonstration Zone issued the first batch of unmanned delivery vehicle vehicle codes in China, and for the first time gave the corresponding right of way to unmanned delivery vehicles, and Baidu Apollo ecological partner Neolithic became the first batch of licensed enterprises. With the implementation of this policy, it means that Neolithic is the first in China to realize the "licensed post" of unmanned delivery vehicles.
In August, Baidu Apollo's new generation of self-driving minibus Apollon II was officially released. It is understood that Apollo II. integrates Apollo's latest generation of self-developed automatic driving computing platform and sensor system, and achieves interoperability with Apollo Robotaxi, and the computing power of the core computing unit is tripled to 372Tops, which is the same as the Computing Power of Robotaxi.
Hitch a ride on the "Car Road Synergy" bus, there is a further road
The industry knows that there are two routes to achieve automatic driving, that is, bicycle intelligence and vehicle-road collaboration. It should be known that in the initial stage of automatic driving, bicycle intelligence has the upper hand, and in the past two years, the car road has come to the front of the stage.
In September 2019, the Autonomous Driving Working Committee and the Autonomous Driving Standardization Working Committee of the China Highway and Highway Society released the "Report on the Hierarchical Definition and Interpretation of Intelligent And Connected Road Systems", which set the tone for the route of automatic driving - "smart car" + "smart road".
As an early player of bicycle intelligence, Baidu's attack on the car-road synergy track is also one step ahead. In other words, Baidu does not take sides in front of the technical route and continues to continue the "I want it all" style.
Three months after the automatic driving route was set, Baidu Apollo made organizational adjustments, in addition to upgrading the original automatic driving, Internet of Vehicles and other business groups, adding an intelligent transportation business group. At that time, according to the official introduction, the intelligent transportation business group focused on the intelligent upgrading of transportation infrastructure, providing cities with next-generation intelligent transportation solutions such as vehicle-road coordination system and regional intelligent signal optimization system.
The overall architecture of Baidu's "ACE Traffic Engine"; Image source: Baidu
In April 2020, Baidu Apollo released the "ACE Traffic Engine", disclosing Apollo's intelligent transportation solution for the first time and systematically interpreting Baidu's business plan in vehicle-road collaboration. This means that Baidu's chess game has officially come to "vehicle-road coordination".
Specifically, Baidu's "ACE Traffic Engine" adopts a "1+2+N" system architecture, that is, "a big digital base, two smart engines, and N big application ecology". Among them, a large digital base refers to digital transportation infrastructure such as "car, road, cloud, and map", including Xiaodu vehicle OS, flying propeller, Baidu intelligent cloud, and Baidu map. The two intelligent engines are the Apollo autonomous driving engine and the vehicle-road synergy engine. N major application ecology, including intelligent information control, intelligent parking, traffic governance, intelligent bus, intelligent freight, intelligent car connection, intelligent rental, autonomous parking and park species.
Up to now, the intelligent transportation engine has landed in dozens of cities, and Baidu has thus become one of the few enterprises in China that can simultaneously power cars and urban transportation. However, it is worth noting that domestic technology companies have sprung up, and companies similar to Baidu's route have also followed up and even surpassed. Huawei, for example, whether it is autonomous driving or vehicle-to-road collaboration, or hardware, software and 5G, has the capital to compete with any other player.
Admittedly, the cake in the field of autonomous driving is big enough, but the competition is also fierce enough. Even if Baidu has the advantage of experience and technical strength, it still needs to make more efforts to stand out in BATH and obtain a more eye-catching report card.