On December 27, Baidu Create 2021 (Baidu AI Developer Conference) was held, and Baidu founder, chairman and CEO Robin Li once again put forward the view that "intelligent transportation can make China's first-tier cities no longer need to restrict purchases and traffic within 5 years, and basically solve the problem of congestion within 10 years", which has attracted widespread attention.
It is not surprising that Robin Li is so "partial" to intelligent transportation. Transportation is related to the national economy and people's livelihood, and more importantly, transportation has become one of the most important application scenarios of AI, whether it is automatic driving or intelligent transportation, it shows the economic, social and industrial value that is visible to the naked eye. In recent years, a number of competent units have repeatedly emphasized the development of intelligent transportation in various guidance documents, and the 14th Five-Year Plan outline released in 2021 proposes to "accelerate digital development and build a digital China", and frames the specific scope of the ten major digital application scenarios as an important guide for building a digital society, improving the level of government digital governance, and creating a good digital ecology.
According to the data released by the China Intelligent Transportation Association, combined with the intelligence of various segments of transportation industries, from 2011 to 2020, the total scale of China's intelligent transportation market increased from 42 billion yuan to 165.8 billion yuan, showing a clear upward trend, with an annualized growth rate of nearly 20%. In this way, in fact, the unmanned vehicle that the public is highly concerned about is automatic driving, which is only the tip of the iceberg of intelligent transportation.
In the past few years, "smart cars" have become the key to the industry to overcome, from the perspective of new trends in the industry, the "smart road" will become a new round of focus of the industry, and more and more technology giants are entering this industry. Baidu, which has already come out on top in unmanned vehicles, has also already placed heavy bets in the field of intelligent transportation, and officially released the AIR (AI Road) intelligent road system at the Create Conference on December 27, which is a "killer weapon" to win the intelligent transportation war.
Vehicle-to-road coordination of technical routes to accelerate the process of traffic intelligence
The AIR intelligent road system is positioned as an "automatic driving system" at the end of the road. Different from the traditional road information equipment that emphasizes the combination of software and hardware, it is a set of self-learning and evolutionary systems that can be driven by data, adopting the same-source architecture as the autopilot system, based on sensors such as lidar, millimeter wave radar, fisheye cameras, cameras and other sensors similar to the autopilot of the car, the calculation unit, software and cloud system, and the application of L4 level automatic driving level perception prediction planning technology, so that the roadside traffic infrastructure can achieve digital, intelligent and networked upgrades.
With the blessing of the AIR intelligent road system, automatic driving will land faster. As we all know, the autonomous driving industry has two routes:
One is the "bicycle intelligence" route represented by Waymo and other European and American players, the problem it faces is that the cost is unattainable, the technical realization is too difficult, and what problems it will face in extreme scenarios such as complex weather and road conditions are not yet known; the other is the bicycle intelligence + vehicle-road coordination (V2X) route represented by Baidu Apollo, so that roads, cars and people are linked to form an intelligent transportation network, this route avoids the problem of bicycle intelligence, in China, which is known as "infrastructure mania", More and more cities are beginning to build roads compatible with autonomous driving, intelligently upgrading existing transportation facilities, and the AIR intelligent road system is to provide cities with the ability to "upgrade the intelligent transportation system OTA".
Vehicle-to-road collaboration with the air intelligent road system will make autonomous driving safer and the cost will be significantly reduced.
The visual distance of the vehicle-side sensor is relatively short, which is greatly affected by complex road conditions and bad weather, and the "car + roadside God perspective perception" can perceive the road surface abnormalities faster and more completely, such as the AIR intelligent road system can fully pay attention to all road risk factors, detect and warn 20+ typical dangerous events such as vehicle retrograde, and effectively reduce 90% of traffic accidents through collaborative active operation management. Smart road + smart car to automatic driving with double insurance, will greatly improve safety. In addition, when the AIR intelligent road system is popularized, the vehicle-side sensors can be reduced and safe automatic driving can be realized, which will greatly reduce the cost of autonomous vehicles and bicycles, thereby accelerating the large-scale commercial process of autonomous driving.
AIR intelligent road system and Apollo automatic driving system synergy, can significantly improve the safety level of automatic driving, reduce the overall cost of automatic driving landing, but automatic driving is only one of its application scenarios, its biggest application imagination space is: to drive the entire traffic system intelligent, and then change the traffic itself.
AIR intelligent road system to "open the eyes" of urban traffic
The core mission of the AIR intelligent road system is to optimize the efficiency of the transportation system, significantly alleviate congestion, and even cure the urban disease of congestion.
Bicycle intelligence focuses only on its own efficiency. When human drivers go to "congestion" when they are stuck in traffic, it will only exacerbate congestion, and similarly, even if all cars become unmanned vehicles, there is no unified scheduling management, and the congestion problem cannot be solved. To be fundamentally solved, the congestion problem needs to be uniformly deployed by the cloud system to achieve the optimal global efficiency, which is another core value of the AIR intelligent road system.
After networking, digitizing and intelligentizing the road infrastructure, based on the real-time traffic data of full perception, the AIR intelligent road system can realize intelligent information control from single-point optimization to trunk line optimization to regional optimization, from traffic flow section detection, fixed timing information control upgrade to intelligent intersection holographic perception, real-time traffic flow simulation prediction, to achieve city-level global optimization control, through traffic light information release, green wave speed suggestion, intersection safety information reminder, traffic event release and other means of active and accurate guidance. Greatly reduce the congestion and delay of urban roads, and the previous practice shows that it can improve the urban traffic efficiency by an average of 30%, and achieve the best traffic efficiency at the city level.
Note that air smart road systems affect not just autonomous vehicles in smart traffic, but can affect all traffic participants, including ordinary vehicles. In the information control traffic management scenario, the traffic management department can combine labor + machine, based on the intelligent traffic system driven by the AIR intelligent road system, to carry out road planning and design, vehicle diversion guidance, accurate reminder of accidents, tidal lane control, etc., to maximize the utilization rate of urban roads and reduce the degree of urban congestion.
Since urban congestion is largely caused by parking difficulties, random parking, and cruising to find parking spaces, the AIR intelligent road system provides services such as roadside sensorless parking and autonomous parking in the yard library through 3D holographic perception and other technologies, making it more convenient for users to travel and park, while alleviating road pressure.
In fact, the concept of intelligent transportation has long emerged, and all places are exploring the intelligence of transportation infrastructure, but on the whole, the "patch upgrade" model of the scattered pilot is adopted, and the support of the unified underlying intelligent system is lacking. With the support of the underlying intelligent system provided by the AIR intelligent road system, the intelligentization, digitization and networking of transportation infrastructure can accelerate the pace, and the system-driven traffic intelligence will be more conducive to global planning, unified upgrading, scale application, and rapid replication.
It can be seen that the AIR intelligent road system is equivalent to the "sky eye" of urban transportation, which will help the city build an overall intelligent transportation system, make the urban traffic efficiency optimal, and make it more convenient for citizens to travel.
Baidu grinds a sword for eight years
Why has there been no intelligent system for intelligent transportation infrastructure in the industry before? Because it's not just a technical issue. Baidu AIR intelligent road system is not made from 0 to 1, but the result of Baidu's years of practice in vehicle-road collaboration, road intelligence and traffic intelligence.
Since the launch of the unmanned vehicle project in 2013, Baidu has been laying out unmanned vehicles for more than 8 years, and it launched the autonomous driving open platform Apollo more than 4 years ago. Unlike European and American players who only do bicycle intelligence, Baidu decided early on to explore the seemingly more difficult route of vehicle-road collaboration, which laid the groundwork for its launch of the AIR intelligent road system today.
In 2016, Baidu set up a team to develop V2X vehicle-road collaboration technology, forming a full-stack systematic research and development capability, covering three major scenarios of highways, parking lots and urban roads, and in 2019, it established an intelligent transportation business group and an open source vehicle-road collaboration program, and has been deeply cultivating on the "smart road" for at least 5 years so far:
Technically, Baidu applies the Apollo autopilot technology to the road end, that is to say, the AIR intelligent road system applies the same technology as the Apollo unmanned vehicle, such as various radars, and the underlying algorithm;
In practice, Baidu intelligent transportation has been applied in nearly 50 cities such as Beijing, Guangzhou, Baoding, Chongqing, Changsha and Yangquan. In November last year, the Baoding Public Security Traffic Police and Baidu jointly built the AI traffic management brain put into use, thanks to the traffic management brain, Baoding City's overall traffic travel efficiency increased by 20-30%, these projects landed Baidu accumulated a wealth of experience in traffic intelligence.
At the Baidu Create 2021 conference, Robin Li said that intelligent transportation can help reduce road traffic safety accidents by 90% and solve the problem of congestion in the system. "Experimental data proves that when all intersections in a city are intelligent and city-level regional information control optimization, the traffic efficiency can be improved by 15%-30%."
It is precisely because it has always insisted on the layout of the "smart road" that Baidu's pattern today is more grand than that of many autonomous vehicle players: because of the smart road, Baidu can make the "smart car" safer, lower cost, and faster landing; it can also eat the bigger cake of intelligent transportation.
The "road of wisdom" has always been valued by Baidu, and it can also be seen from the book "Intelligent Transportation: Major Changes Affecting Mankind in the Next 10-40 Years" (hereinafter referred to as "Intelligent Transportation") written by Robin Li, founder, chairman and CEO of Baidu in 2021.
Robin Li's original intention in writing "Intelligent Transportation" is to "find the best solution for traffic safety, traffic congestion, carbon emissions and other issues, so that each of us can travel safely, efficiently, economically and greenly." "This book is not a simple science popularization of intelligent transportation, but a discussion of Baidu's experience in strategically laying out autonomous driving and intelligent transportation over the years." In the book, Robin Li systematically explains the concept of intelligent transportation system for the next 10-40 years, and proposes six innovative concepts and models, of which a large number of pages reflect the great importance attached to the "smart road".
For example, Robin Li believes: "Vehicle-road collaborative automatic driving is a 'Chinese model' with Chinese advantages. "The book "Intelligent Transportation" vividly points out that vehicle-road collaboration is like a street lamp, while bicycle intelligence is like a lamp, and vehicle-road collaboration can greatly improve the safety of unmanned vehicles and accelerate the process of large-scale commercialization of autonomous driving." The adaptive traffic signal control system "will be the infrastructure for the urban intelligent transportation system of the future." With the rapid development of artificial intelligence technology and the gradual popularization of infrastructure such as 5G, it is gradually possible to use intelligent algorithms to predict based on real-time traffic flow information, and coordinate intelligent and adaptive signal light adjustment at all intersections and lanes in the city.
Baidu has always attached importance to the "smart road" has also been in the strategic layout, and now the release of air intelligent road system, it is based on the previous years in the practice of vehicle-road collaboration, intelligent transportation, it can be said that the AIR intelligent road system is Baidu from the landing experience of nearly 50 cities to summarize the intelligent transportation system-level technology, capabilities and experience, both from the accumulation of automatic driving technology, will also be in-depth coordination with the automatic driving business, to create a "Chinese model" for urban transportation, Make Baidu an important player in the new infrastructure of intelligent transportation.
The "China Model" to Help a Transportation Powerhouse
"Transportation" is related to the lifeblood of the country's livelihood, economy and security, "transportation power" has also been an important national policy in China, as early as 2019, the CPC Central Committee and the State Council officially issued the "Outline for the Construction of a Transportation Powerhouse", depicting the blueprint for the construction of a transportation power in China from now to the middle of this century, the "Outline" clarifies the key tasks and safeguard measures for the construction of a transportation power, and puts forward the task of "scientific and technological innovation is full of vitality and wisdom to lead" in the field of transportation science and technology, requiring the strengthening of cutting-edge scientific and technological research and development, Vigorously develop intelligent transportation and improve the mechanism of scientific and technological innovation.
The "Outline for the Construction of a Transportation Powerhouse" clearly states that it is necessary to comprehensively improve the intelligent level of urban transportation infrastructure, and in the part of "vigorous development of scientific and technological innovation and wisdom leadership", it is proposed to "vigorously develop intelligent transportation", strengthen the research and development of cutting-edge key technologies, including strengthening the coordinated operation and service technology of regional comprehensive transportation networks, and the coordinated control technology of urban transportation, etc., and at the same time improve the scientific and technological innovation mechanism, "Establish a technological innovation mechanism with enterprises as the main body and the deep integration of production, education and research, and encourage all kinds of innovative entities in the transportation industry to establish innovation alliances." Establish a key core technology research mechanism. ”
Intelligent transportation is the key direction of a transportation power. The "14th Five-Year Plan" period is an important stage in accelerating the construction of a transportation power, the 14th Five-Year Plan outline released in 2021 puts forward a number of task requirements for transportation, deploys the construction of a transportation power, under the strategy of "new infrastructure" and "construction of digital China", intelligent transportation is more valued, intelligent technology triggers transportation reform, and will also be an important opportunity for China to upgrade from a transportation power to a transportation power.
Traffic intelligence is the first in road intelligence, Baidu released the AIR intelligent road system at this time, which not only reflects the development direction of the main body of the enterprise participating in smart transportation, but also conforms to the key direction of "improving the intelligent level of urban transportation infrastructure" and other transportation powers. Baidu AIR intelligent road system will also contribute to the "China model" of a transportation powerhouse.
Making all transportation infrastructure intelligent is also the vision of Baidu's intelligent transportation strategy. In this book, the future core of intelligent transportation is an intelligent transportation system built by new technologies, new concepts, and new models, which can reduce traffic safety accidents by 90%.
To achieve real intelligent transportation, only rely on the intelligence of the car is far from enough, systematic upgrading, large-scale investment, long-term construction will be the inevitable path, the road and other transportation infrastructure intelligence, can not be achieved overnight.
Write at the end:
There are more and more commercial cases of autonomous driving in 2021, but there are still many gaps to be crossed to make autonomous vehicles change from the cool concept of the past to people's daily lives. Now it seems that building a "smart road" after a "smart car" will not only accelerate the commercial process of autonomous vehicles, but also completely transform the entire transportation system, greatly improve traffic efficiency, and make travel safer, more convenient and efficient. In the critical period of transportation power, intelligent transportation will also become a key opportunity for the rise of China's transportation. Baidu AIR intelligent road system, born at the right time.