A "traffic evolution" is coming: What upgrades will 5G bring to vehicle networking and autonomous driving?

• Note: This article is reproduced from Tencent Technology, author: Li Yan, Senior Director of Technical Standards at Qualcomm

2019 is regarded as the "first year of 5G" globally, and China's 5G development has also entered the sprint stage. As a new generation of mobile communication technology, 5G will become a unified connection architecture to support future innovation and give new impetus to economic growth.

With the advent of 5G commercialization, people are expecting this innovative technology to bring a new look to many traditional fields. For example, for the century-old automobile industry, this will mean more acceleration and breakthroughs in autonomous driving and vehicle networking technologies, and even open up more room for imagination for the entire industry.

First of all, let us understand the concept of "Internet of Vehicles", which is also called V2X (Vehicle to Everything). It is the interconnection between cars and everything, including communication between cars (V2V), cars and infrastructure (V2I), cars and pedestrians (V2P) and networks (V2N).

The Internet of Vehicles is like a WeChat group that includes cars, roadside facilities such as traffic lights, pedestrians, and cloud participants. Every participant in the group can share his or her information with other participants in real time, recognize each other's location and driving intentions, and notify traffic information such as traffic lights, etc., to assist vehicles in the group in perceiving the road and support vehicle automation.

Traditional V2X technology has been around for a long time. As a derivative technology of the early 802.11a technology, it was conceived in the first few years of the 21st century. However, traditional V2X technology has obvious limitations. The main ones are as follows:

- First, it lacks a long-term wireless technology evolution roadmap. As vehicles become more and more connected, the automotive industry has embraced 4G LTE technology and is moving towards a 5G future. This legacy technology has become out of touch with the automotive industry’s future vision and development direction.

- Secondly, there is a lack of motivation for investment in traditional V2X technology. For automakers, a one-time investment is not cost-effective, and traditional technology will be unsustainable over time; as far as road infrastructure is concerned, traditional V2X technology cannot produce synergy with other widely deployed and continuously innovative wireless technologies, and it is difficult for transportation departments at all levels with tight funds to invest in it and promote infrastructure upgrades.

- Finally, traditional V2X technology did not foresee the acceleration of the global 5G transformation in 2019-2020. It is precisely because of the acceleration of global 5G technology development that the investment direction of automobile manufacturers and road infrastructure managers has changed.

In view of the shortcomings of traditional V2X technology and taking full advantage of the industrial scale of cellular mobile communications, the global mobile communications standardization organization 3GPP defined C-V2X (cellular vehicle-to-everything) technology in the R14 standard version. Since the birth of C-V2X technology in 2016, most participants in the automotive industry have adopted C-V2X, a technology that can comprehensively solve safety and efficiency problems. It is the best technology that is suitable for future development and in line with the global transformation path.

The global wireless deployment of C-V2X based on 3GPP wireless standards is in line with the 5G evolution path and will give full play to the scale effect from the terminal and infrastructure levels. C-V2X can be integrated into the in-vehicle information processing wireless module, so the incremental cost problem of automobile manufacturers is also solved. The deployment of road infrastructure can produce synergy with the large-scale deployment of 5G, which saves a lot of expenditure for governments at all levels. Compared with traditional technologies, C-V2X also has technical advantages and can provide more than twice the communication range and reliability.

Currently, autonomous driving technology is developing rapidly in China and the United States. By the end of 2018, the California government had issued autonomous driving test licenses to 60 companies, and local governments in China had also issued test licenses to 24 companies.

In the latest annual autonomous driving competitiveness rankings released by the famous research organization Navigant Research, most of the leading companies adopt a single-vehicle intelligence approach, that is, the vehicle's perception of the environment and driving decisions are completed through on-board sensors and computing processing units. This can achieve faster progress in a short period of time on urban roads with relatively complete traffic facilities or on highways with relatively simple working conditions.

However, relying solely on the intelligence of a single vehicle will have significant limitations. For example, for roads with serious traffic infrastructure defects and irregular deployment, or highways with heavy traffic and high speeds, it is difficult for a single vehicle to perceive and make real-time decisions in complex road environments. However, with the development and application of 5G, these limitations have been broken for self-driving cars.

Thanks to the continuous advancement of 5G technology and applications in the Chinese automotive industry chain, the industry's autonomous driving research field generally believes that autonomous driving requires not only smart cars, but also smart roads. Using the low latency, high reliability, high speed and large capacity of 5G technology, the Internet of Vehicles can not only help vehicles communicate about their location, speed, driving direction and driving intentions, but also use roadside facilities to assist vehicles in perceiving the environment. For example, a vehicle may not be able to accurately judge traffic lights using its own camera, and may then violate traffic rules by running a red light. However, using the V2I technology of the Internet of Vehicles, traffic lights send light signals to surrounding vehicles in the form of wireless signals, ensuring that autonomous vehicles accurately understand the status of traffic lights.

Not only that, traffic lights can also broadcast the time of the next signal change, and even the signal status of other adjacent intersections in the future. Autonomous driving vehicles can accurately optimize their travel speed and route based on this, and choose a route with the fewest red lights and the fastest travel, which not only optimizes traffic but also reduces carbon emissions.

Another example is traffic optimization and warning of pedestrians crossing the road at intersections. Today, collisions between pedestrians/bicycles crossing the intersection and vehicles often occur on the road, especially left-turning vehicles. Due to obstructed vision, drivers and on-board sensors often cannot observe pedestrians crossing the intersection. One solution is to monitor pedestrians in the intersection by installing radars and cameras at the intersection. If pedestrians are detected on the zebra crossing and in the intersection, and the pedestrians are on the vehicle's route, the roadside facilities (RSU) can immediately notify the vehicles about to turn or go straight of the detected situation to pay attention and avoid accidents.

5G and Internet of Vehicles technologies are necessary technical guarantees for autonomous driving. By choosing the correct technical route of "smart cars + smart roads" and giving full play to the technical advantages of 5G, we have reason to believe that 5G technology will play a huge role in the field of autonomous driving.

Autonomous driving of cars includes three levels: perception/cognition, decision-making and execution, and all three levels can be enhanced using 5G mobile communication technology.

First, let's look at the perception/cognition level. Perception/cognition is to let the car and the driver know "where I am, what is around me, and whether I am in danger." Therefore, high-precision positioning is required first. For autonomous driving, even sub-meter positioning is required. Traditional satellite positioning (GNSS) cannot meet the positioning accuracy required for vehicle automation, so it is necessary to use network-assisted differential positioning technology and combine it with track deduction technology to improve positioning accuracy.

Take Qualcomm’s Vision Enhanced Precise Positioning (VEPP) technology as an example. It integrates multiple automotive sensors, including the Global Navigation Satellite System (GNSS), cameras, inertial measurement units (IMUs) and wheel sensors to provide more accurate global vehicle positioning.

VEPP is a combination of a telematics control unit and a high-level positioning device, which achieves lane-level accuracy with an accuracy of less than 1 meter. After accurately grasping our own position, we also need a lane-level high-precision map that supports autonomous driving to know the position of the vehicle on the road. By taking advantage of the large bandwidth of 5G mobile communications, vehicles can obtain the latest high-precision maps in real time.

Secondly, the vehicle needs to understand the surrounding road participants and obstacles. Autonomous vehicles are often equipped with multiple advanced sensors, including cameras, radars, and even lidars to sense surrounding objects. However, these expensive sensors can only detect objects within the line of sight and are also sensitive to climatic conditions. For example, cameras and lidars will fail in fog and haze weather. By using V2V vehicle-to-vehicle communication technology, the vehicle's perception range can be expanded beyond the line of sight, such as the position of several vehicles in front and behind, and even their emergency braking status information, so that the road conditions can be judged in advance and evasive measures can be taken as soon as possible to avoid rear-end collisions. At the same time, using V2I communication, vehicles can also obtain information such as traffic lights and pedestrians at intersections from road infrastructure to form a complete perception of the road environment.

5G technology can not only improve the environmental perception capabilities of autonomous vehicles, but also use wireless connections between vehicles to allow multiple vehicles to make collaborative decisions and plan actions reasonably. For example, when a vehicle on the inner lane of a highway needs to leave the highway, it can negotiate with surrounding vehicles through vehicle-to-vehicle communication and ask surrounding vehicles to give way so that it can change lanes to the outer lane and leave the highway.

As mentioned above, L4/L5 is only theoretically feasible if it relies solely on single-vehicle intelligence. It is an almost impossible task to teach machines to accurately identify all signs, signals, and road participants on the road, and it is also an unbearable cost to teach machines to handle all possible incidents. Only through vehicle-to-vehicle/vehicle-road collaboration based on 5G technology and by giving full play to the capabilities of the infrastructure can autonomous vehicles perceive the driving environment and make driving decisions at a lower price and cost.

In addition, from the perspective of artificial intelligence, machines cannot completely replace human decision-making. In some occasional complex scenarios, humans are required to participate in decision-making, that is, "man in the loop". However, for self-driving vehicles, the passengers in the car may not be able to drive the car and cannot participate in decision-making.

But with 5G technology, we can allow the personnel at the remote service center to participate in the decision-making. For example, in some special scenarios, the machine cannot complete the driving and causes disgrace, just like when we call the rescue center when the vehicle breaks down today, the vehicle can use the 5G network to call the remote service center, and use 5G's low latency and large bandwidth to transmit the real-time images and sensor information on the scene to the professionals at the service center. The professionals can remotely control the vehicle to drive away from the complex road conditions in a virtual reality scenario until the driving system can take over the vehicle again.

Of course, we are not saying that only highly automated vehicles such as L4/L5 require 5G technology. In fact, Internet of Vehicles technology is also of great help to low-level automation centered on human driving. For example, the many perception technologies mentioned above can be applied to current vehicles to provide drivers with driving assistance (ADAS), such as forward collision warning and emergency automatic braking based on the fusion of vehicle-mounted radar and V2V technology.

At present, with the strong promotion of the Ministry of Industry and Information Technology and the Ministry of Transport, the Internet of Vehicles based on the 3GPP R14 version has basically completed the technical preparations for mass production. Of course, there are still some policy and technical issues that need to be resolved urgently before large-scale commercial use.

First, it is about information security and privacy. For the general public, in addition to paying attention to the many conveniences and road safety brought to us by the Internet of Vehicles, they are also more concerned about the information security and privacy issues designed by the Internet of Vehicles. In order to protect the privacy of car owners and improve the security factor of the network, C-V2X has designed an anonymous key system, which periodically changes the identification and information signature. For this purpose, hundreds or even thousands of certificates need to be installed in the vehicle. Because it involves information security issues, it is urgent to cooperate with multiple relevant national departments to formulate a certificate management and distribution system suitable for China's national conditions.

In addition, in terms of C-V2X information security, China's self-developed SM2/3 encryption algorithm is required. This algorithm has been widely used in banking and other systems and has a certain industrial foundation. However, the processing power of existing domestic chips that support the SM2/3 algorithm cannot meet the requirements of thousands of signatures and signature verifications per second required by C-V2X, and most domestic companies engaged in security chip development lack experience in developing automotive-grade products. Although foreign chip manufacturers have obtained the SM2/3 algorithm from the national standard system, they have been slow to come up with product plans that support C-V2X security specifications because they do not fully understand the certification links and related certification risks involved in the national encryption algorithm. Therefore, it is necessary to further promote it from the perspective of policy and funding. It is expected that domestic and foreign companies engaged in security chip design and development will quickly come up with automotive-grade chips with high processing power that support the national encryption SM2/3 algorithm.

C-V2X based on 3GPP R14 can enhance the perception of the environment by autonomous vehicles and provide more accurate driving assistance information for low-level autonomous driving. The collaborative decision-making required for L4/L5 advanced autonomous driving will place higher requirements on V2X communication, requiring V2X to provide higher bandwidth, shorter latency, and most importantly, higher reliability (>99%).

In March 2019, 3GPP approved the standardization project and planned to develop 5G V2X technology based on 5G new air interface in the R16 version, using the high bandwidth, low latency and high reliability of 5G new air interface to further enhance the capabilities of C-V2X technology and support high-level autonomous driving technology. The R16 version of the C-V2X standard is scheduled to be completed by the end of 2019. It is believed that after the standard is completed, chip manufacturers will also launch chips that comply with the R16 version as soon as possible to cooperate with car companies in the development and testing of L4/L5 high-level autonomous driving vehicles.

Autonomous driving technology can free us from the heavy tasks of driving, making our journey safer and more environmentally friendly. 5G Internet of Vehicles will greatly enhance the perception capabilities of autonomous vehicles and promote the early mass production of autonomous vehicles. More advanced autonomous driving requires extensive collaboration between vehicles and roads, and 5G technology is an irreplaceable and necessary means for autonomous driving.

Although 5G application in the Internet of Vehicles has not yet been truly realized, it is on the way to realization. After 3GPP released the R14 version of the C-V2X standard in June 2017, the industry quickly launched the development of C-V2X products and related testing and verification work. In September of that year, Qualcomm released the 9150 C-V2X chipset based on the 3GPP R14 version and carried out extensive testing around the world based on this chipset.

In November 2018, the Ministry of Industry and Information Technology issued the "Management Regulations for the Use of the 5905-5925MHz Frequency Band for Direct Communication of Internet of Vehicles (5905-5925MHz for Intelligent Connected Vehicles) (Interim)", which planned a dedicated frequency resource with a total bandwidth of 20MHz for C-V2X (direct communication between vehicles, vehicles and people, and vehicles and roads) direct communication technology for intelligent connected vehicles. At the same time, it made regulations on the management of related frequencies, stations, equipment, and interference coordination, making China the first country in the world to clearly allocate dedicated frequencies for C-V2X technology. The division of this frequency band has completely cleared the policy uncertainty in the promotion of Internet of Vehicles, and will greatly promote car companies to invest in the development of mass-produced vehicles with C-V2X technology.

During this year's Consumer Electronics Show (CES), Ford Motor Company announced that starting in 2022, every new car and truck sold in the United States will have C-V2X technology built into it; and recently announced that it will mass-produce the first model equipped with C-V2X technology in China in 2021.

During the Mobile World Congress in February this year, Geely announced plans to release its first batch of mass-produced models supporting 5G and C-V2X in 2021. Some domestic automakers have also actively followed up and are considering supporting C-V2X technology first in some of their mass-produced models within two years.

On March 28, Minister Miao Wei of the Ministry of Industry and Information Technology attended the Boao Forum for Asia Annual Conference 2019 and stated that the Internet of Vehicles represented by driverless cars will be an important application of 5G. The Ministry of Industry and Information Technology is working with the Ministry of Transport to actively promote the digital and intelligent transformation of China's highway system and use 5G technology to give full play to the advantages of vehicle-road collaboration.

We have reason to believe that in the next two years, with the construction of 5G networks, most cities and highways in China will use C-V2X technology to upgrade their transportation infrastructure. Vehicles with C-V2X functions that will be mass-produced from 2020 will be the first to experience the convenient transportation, safe driving tips and more optimized energy conservation and emission reduction brought by C-V2X technology. The combination of 5G and emerging technologies such as AI will surely stimulate a wave of innovation in the automotive industry and enable the evolution of the entire automotive ecosystem.

JD.com responds to the cancellation of the contract with the 2019 campus recruitment; experts explain the three major difficulties in publishing the black hole photo

Just now, EHT released the first real black hole image of mankind

Pinduoduo responds to Apple's suspension of supply; WeChat opens advertising @ friends function; Huawei is rumored to only sell 5G chips to Apple

A conversation with Roadstar investors: The death of an autonomous driving company

Google's new AI ethics committee has collapsed, is its smart city project also going to fail?

Alibaba pays more than 140 million yuan in taxes every day on average; Final judgment on Samsung Note7 explosion case