Is “vehicle-road collaboration” the ultimate answer to the implementation of autonomous driving technology?

This morning when I was taking a taxi, the driver suddenly pointed to me and said, "Look at that car, it's driving very slowly and very orderly, it must be an autonomous car." I passed it and saw that it was indeed the case. It can be said that the existing mass-produced car autonomous driving system is still very "inhumane", meaning that you can tell at a glance that it is a machine driving, not a person.

In recent years, there have been endless discussions about autonomous driving. Most of the current mass-produced cars are equipped with L2 and L2+, L2.5 and other autonomous driving systems, but in fact, we can now experience higher levels of autonomous driving technology. For example, Waymo, a subsidiary of Google's parent company Alphabet, has been able to operate Robotaxi without a safety officer.

In February this year, the National Development and Reform Commission and 11 other ministries jointly issued the "Intelligent Vehicle Innovation and Development Strategy", which identified "vehicle-road collaboration" as an important goal for future development and also put this term in the forefront.

Indeed, with the support of 5G and new infrastructure, China is most likely to achieve vehicle-road collaboration.

The internetization, automation and intelligence of automobiles have become the established direction of automobile development, and cars that can achieve autonomous driving are considered to have artificial intelligence (AI) capabilities, which can automatically drive, perceive their surroundings, identify objects nearby, and perform reasoning and decision-making.

Vehicle-road collaboration requires vehicles to use the 5G-V2X network and IoT services to interact with other vehicles on the road, even pedestrians and non-motor vehicles.

Simply put, vehicle-road collaboration requires not only "smart cars" but also "smart roads".

Smart car

"In terms of autonomous driving, which is mainly based on single-vehicle intelligence, we are two years behind the United States in technology."

Chen Yinzi, vice president of Shanghai Zhongzhixing Technology, believes that it is difficult to surpass others in a short period of time with the technology accumulated through a large number of tests, but we can still adopt the most suitable way for China to overtake. In fact, vehicle-road collaboration can greatly reduce the computing power requirements of the car itself, but the single car still needs to have a certain perception ability.

Unmanned driving technology mainly uses lidar, millimeter-wave radar, high-definition cameras and sensors to capture scene information around the vehicle, and outputs instructions to the computer through AI artificial intelligence algorithms to control the vehicle's steering, braking and speed, so that the vehicle can drive safely.

While most manufacturers are still investing in the research of single-vehicle intelligence, CIRI already had mature L4 autonomous driving last year and turned to the direction of vehicle-road collaboration.

In the current booming automotive industry, many people hope that autonomous driving can improve traffic congestion and reduce traffic accidents. In fact, the detection range of sensors on a single vehicle is limited, and vehicles and pedestrians blocked by obstacles at close range cannot be identified. The realization of single-vehicle intelligence is more about liberating the driver. Once encountering bad weather or sudden accidents, the overall vehicle-road collaboration becomes more important.

Sensors installed on both sides of the road can allow vehicles to predict the road conditions several kilometers ahead, as well as vehicles and pedestrians blocked by obstacles around them, to achieve higher safety. Especially in bad weather such as rain, snow and fog, roadside sensors can also make up for the limitations of single-vehicle sensors in detection capabilities and further improve safety.

At the same time, in the vehicle-road collaborative system, vehicles and roadside equipment can also interact with the "cloud brain" of the urban transportation system in real time, and the urban cloud will coordinate the movement of vehicles, greatly improving traffic efficiency. For example, when encountering a traffic light, waiting vehicles can start at the same time, alleviating or even eliminating the tidal phenomenon. In addition, the urban cloud can also plan lanes more reasonably, greatly reducing congestion.

Let's take an extreme example. An unmanned smart car is driving on the road. When it passes under a bridge, a street lamp falls from the bridge. If it is a single smart car, it can recognize the street lamp after it falls and go around it. But if it happens to hit your car, no matter how smart the car is, it can't do anything. If it is a car-road collaboration, the roadside sensing unit will notify the surrounding vehicles at the same time when the street lamp falls. Your vehicle and the following car behind you will automatically stop under the bridge. When the street lamp falls, you just need to go around it safely.

The Path of Wisdom

In order to achieve environmental perception and vehicle-road collaboration, the development of road-side equipment is very important: Continental’s current products in the field of smart transportation include sensor systems, processing systems, and network connections to provide a complete set of unified solutions to support the three major elements of smart transportation: perception, decision-making, and interconnection.

Perception link: Continental's high-performance all-weather 77GHz millimeter-wave radar sensor, combined with the self-developed AI intelligent image fusion algorithm traffic camera with independent intellectual property rights, is like installing eyes for the road. The image and radar sensors complement each other very well, and the Doppler effect can achieve all-weather, uninterrupted, efficient and accurate capture of road-side information.

Decision-making and processing link: In addition to installing a clairvoyant on the road, a powerful brain, the "edge computing unit MEC", is also needed to provide computing power, so as to establish an environmental model with the data obtained by multiple sensors and complete the seamless connection of data between sensors, and finally output the results that support decision-making, providing a solid guarantee for the intercommunication of signals and data between sensors.

Network connection and intercommunication link: After obtaining the road-side information, our road-side communication unit will interact with the vehicle-side communication unit through the 4G/5G wireless network, thereby achieving complete integration and docking of the vehicle-side signal and the road-side signal through the background system service architecture, adopting the characteristics of low latency and high coverage.

Continental's environmental sensing hardware system can monitor road conditions around the clock. At the same time, the processing system will integrate various environmental information and signals collected by the sensing system and convert them into key operating factors such as traffic flow information, traffic information, and road conditions. The connected system connects the car and the road, transmitting road and traffic information from the road side to the car, and transmitting vehicle information from the car side to the road.

Gu Wei, head of Continental's Commercial Vehicles and Services Division in China, revealed: "Continental has deployed environmental perception solutions in major application scenarios of China's transportation. Currently, there are some key pilot cities that hope to cooperate or are in negotiation. In the future, we will further increase the research and development of new products and the implementation of projects in the field of intelligent transportation to provide support for China's smart transportation construction."

In the initial realization of vehicle-road collaboration, Continental has established a localized commercial vehicle Internet of Vehicles ecosystem, with safety, cost and efficiency as core concerns.

The repair and maintenance module helps drivers develop fuel-efficient driving habits and provides the fleet with the most timely, accurate and economical repair and maintenance solutions. The intelligent scheduling and autonomous driving modules provide the fleet with the most reasonable transportation routes, the most efficient scheduling plans and the most intelligent management solutions. The vehicle inspection, driving coach and online engineer modules improve the inspection process, supervise and regulate the driver's driving behavior habits, conduct online detection of vehicle faults, and propose solutions to improve the safety of the fleet operation.

Vehicle-Road Collaboration

Now, everyone is crossing the river by feeling the stones. Before the real national standard is unified, the state of letting a hundred flowers bloom can greatly promote the development in this area.

We should seize the opportunity of China's current global leading advantage in 5G technology, combine AI-enabled single-vehicle intelligence with 5G, and equip single-point intelligent vehicles with all-weather, all-scenario, 360-degree "clairvoyance" and a "smart brain" that coordinates the overall situation, to achieve the interconnection and mutual control of all things among vehicles, roads, people, and infrastructure.

In the future, every car will become a receiver, transferor, processor and sender of information, and global coordination will be carried out on a city smart network.

According to China Intelligent Driving, on July 11, 2020, the Technology Innovation Center of China Telecom Corporation Limited, Shanghai Economic and Information Commission, and China Intelligent Driving (Shanghai) Transportation Technology Co., Ltd. formally signed a tripartite cooperation agreement on "Key Technology Research and Industrial Application of Intelligent Vehicle-Road Collaboration".