The implementation of C-V2X will accelerate the process of autonomous driving

In November 2020, the U.S. Federal Communications Commission (FCC) formally voted to allocate the 5.9GHz frequency band to Wi-Fi and C-V2X (Cellular Vehicle-to-Everything, a cellular network-based vehicle networking communication technology). This marks that the United States has officially abandoned DSRC and turned to C-V2X. my country's main promotion of C-V2X has become the only standard for the global vehicle networking.

C-V2X, the only global Internet of Vehicles standard However, smart cities and self-driving cars have not yet come together, and communication, security and power consumption issues still need to be resolved. For this reason, not only China, but also European and American countries are thinking about how to develop smart roads and smart cars. 

1. China takes the lead

According to the C-V2X standardization work plan, combined with the new product development cycle and time to market, it is expected that China will have the foundation for large-scale deployment of C-V2X by 2022. Previously, pilot demonstration work in demonstration areas and pilot areas has been carried out one after another. As the C-V2X deployment work is implemented, it will first bring about massive demands for chips, modules, terminal equipment, management platforms, security certification, etc. At the same time, transportation information manufacturers with strong implementation capabilities will also benefit directly.

The progress of C-V2X is largely related to the system. The Chinese government is very active in investing in infrastructure and plans to install V2X technology facilities on 90% of highways. China is becoming a leader in this field. Europe and Japan will follow closely. Since no one expects the US government to pay for such things, but hopes that the private sector can step up, the US will lag behind the former.

In recent years, the debate between the North American communication standard 802.11p (DSRC, dedicated short-range communication) and C-V2X has been going on. The FCC designated DSRC as the technical standard for its business more than 20 years ago, but DSRC has not been actually deployed, and this critical mid-band spectrum has been basically idle for decades. For this reason, the FCC began to shift from DSRC to C-V2X to accelerate the actual deployment of its services and thus improve automobile safety. 

2. 5G is the only way forward

The political system will play a role in the implementation of V2V (Vehicle to Vehicle) and C-V2X. Therefore, the United States is different from Europe and is also very different from China.

Currently, some innovative American cities are trying to invite autonomous vehicles into their areas through some early deployments. Germany's Volkswagen GTI was equipped with DSRC, and General Motors has provided 4G capabilities for several years, but its use has been quite limited. It is not used to send updated software, but only to help owners who are unable to enter their cars unlock.

C-V2X is tightly coupled with the development of fully autonomous vehicles, and the common denominator in technology is 5G. 3GPP has implemented some specific features in 5G to make it more useful in the automotive sector, including improved reliability and bandwidth, and some added priorities. For example, if someone in the back seat of the car is playing a movie, the data can also be prioritized, which helps safety applications.

C-V2X uses 5G for communications, such as planning routes for autonomous driving. Many OEMs have already committed to deploying V2X in telematics control units (TCUs). Roadside units on the infrastructure side, which communicate with streetlights or parking lots, are also using 5G radios. 

C-V2X utilizes 5G communications The deployment of C-V2X infrastructure depends on supportive regulations, significant investments, and the development of key technologies for autonomous vehicles. As the level of autonomous driving of cars increases, more and more countries are beginning to participate in pilot projects to test C-V2X applications in reality. Many US OEMs have pledged to support this feature in new models. Michigan has established the Office of Future Mobility, which is responsible for the strategic coordination of all travel-related initiatives (including infrastructure); San Francisco, Atlanta, Colorado, Pittsburgh and other places have also launched pilot projects for autonomous vehicles and C-V2X.

Likewise, in Europe, C-V2X trials are underway in many countries. There is broad support from private companies and initiatives such as the EU C-Roads project. 5G readiness testing is also underway in anticipation of widespread 5G rollout in the region. 

3. Vehicle-road collaboration is not enough

Smart city planners and OEMs are working to solve the various problems facing C-V2X, but the two often work at very different paces and with very different technologies—even though the two worlds need to communicate in order to work.

Optimizing traffic in urban areas is critical to reducing energy consumption and accidents and getting emergency vehicles through as quickly as possible. However, achieving these goals is complex and requires a combination of communications, safety, and power technologies that need to be implemented in roads, traffic signals, buildings, and other structures to enable vehicles to communicate with each other and with infrastructure.

C-V2X uses cellular protocols to enable direct communications between vehicles and obstacles such as other vehicles, cyclists, pedestrians and road workers, and to receive safety information from roadside transmitters.

Whatever C-V2X is, it will greatly enhance the capabilities of self-driving cars because it enables the exchange of information at close range. Apart from C-V2X, no sensor can see what's going on around an intersection.

So far, the development of autonomous driving has been vehicle-centric or regulation-centric. But if you have C-V2X in your car, even if you can talk to the city, you can't communicate with other vehicles because the market penetration is still low. So the car side has to catch up. This is a safety consideration.

In addition, cars communicating with intersections through smart traffic signals can also improve travel efficiency, which does not require large-scale market penetration. Another example is emergency service vehicles and first responders. If they can dial into C-V2X themselves, they can get green lights all the way and "control" the entire road, not just in the near field. This is an application of wireless wide area network and short-range network. The response is quite fast and the benefits are immediate.

A smart city is not a set of related static infrastructures, but a dynamic collection of capabilities, so any road user is part of a smart city, including pedestrians and vehicles.

While both C-V2X and autonomous driving are being developed, their timelines for implementation may be very different. Even if a vehicle using LiDAR comes to an intersection, it does not mean it is a L4 or L5 autonomous vehicle, it is likely just L2. In fact, with C-V2X, many vehicles can provide a higher level of driving capabilities, and it will not be as expensive as L4 or L5.

Commercial vehicles are likely to be the best scenario for the implementation of vehicle-road collaboration, although the infrastructure is different. These facilities will be deployed in transportation corridors, and the first use case is to solve the "last mile" site-to-site transportation, especially in highly urbanized areas.

As for the vehicles themselves, how the C-V2X concept will work inside a car remains to be seen. However, power consumption is a tricky consideration from a data processing perspective, no matter where it is. Especially in pure electric vehicles, there is a trade-off between energy consumption and performance.

Additionally, off-the-shelf processors are not suitable for electric vehicles or infrastructure equipment. While it is technically possible to use off-the-shelf hardware in infrastructure (such as roadside signs) running the exact same algorithms for object detection and classification, it is not feasible in terms of power consumption.

An example is an intersection during rush hour, where the roadside unit (RSU) detects more objects, but the processing is computationally demanding. One would like to be able to get all the computing power from solar panels, but this is not possible because the solar panels would have to be the size of a football field. Back in the vehicle, it has to run AI inference algorithms, which require a lot of training and also a lot of processing. All of this consumes a lot of energy. This means that the idea of ​​off-the-shelf hardware components is impractical, especially at a low price.

To be sure, smart cars that can communicate with each other and their environment open up the possibility for safer roads, more efficient travel, and a better driving experience, but adoption will take time.

V2I (vehicle-to-infrastructure) and more broadly V2X require supporting smart infrastructure outside the vehicle. Although pilot projects are already underway, we are still in the early stages of understanding the requirements and capabilities of existing technologies. Large-scale adoption is expected to be at least five years or more away. 

4. Security is a severe challenge

Regarding infrastructure, if someone puts a fake sign on a utility pole, how will the vehicle react? This attack mode has happened before. Tesla's Autopilot system has been successfully "tricked" by radio waves, sound and light many times, and it has started to slow down, brake or turn. Is this dangerous?

The biggest challenge of vehicle-road collaboration is also security. First of all, you must ensure that the car communicating with the vehicle is a trustworthy car, not a fake signal sent by some hacker trying to interfere with the owner's response. This involves public key encryption for mutual authentication between cars. A car will receive hundreds of signals from all the cars around it, which are actual cars from Volkswagen, Mercedes, General Motors, etc. In order to do this, the vehicle needs a root of trust to ensure that no one can fully view your encryption processing and the encryption operations when processing them.