Let you understand the relationship between V2X and autopilot.

Since the development of autonomous driving technology, the ultimate goal of practitioners can be summarized in one sentence, which is to make the thinking and decision-making of vehicles infinitely close to human beings and even surpass human beings in safety when facing complex road information data. This is also taken as a handle by the empty singer-the machine is a machine after all, and its essence has certain limitations, which makes the unmanned vehicle far less "promising" than it sounds.

Therefore, in view of a series of weaknesses of vehicles, witty engineers overcame all difficulties and put forward a set of solutions one after another. Among them, V2X technology known as "self-driving car clairvoyance" has been put on the agenda by major manufacturers. Even if this technology can be commercialized at this stage, Bosch has thought it over for everyone: in APA (Automatic Intelligent Parking) system, V2X is used to solve the "last mile" problem.

As for V2X technology itself, the definition given by the industry is: to realize the information interaction between vehicles and all entities that may affect vehicles, and to complete the perception work through radio wave propagation, that is, wireless communication, so as to reduce accidents, slow down traffic congestion, reduce environmental pollution, and provide other information services. V2X mainly includes vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), vehicle-to-internet (V2N) and vehicle-to-pedestrian (V2P).

To put it simply, V2X means "the car is connected with everything", creating an "over-the-horizon sensor". It is necessary to know the mainstream sensors currently carried by self-driving cars, including cameras, millimeter-wave radars, and lidar. The advantage of V2X, a vehicle-mounted communication technology, is that it can let the vehicle know the surrounding dynamics in real time, detect the traffic flow changes out of sight and send an alarm message to the driver through a low-delay and high-reliability network connection interaction mode.

Although V2X has strict requirements for network connection, once the technology is "on the bus", the control efficiency will be obviously better than that of human drivers. Then the question is, is it necessary for us to "crash" with all kinds of physical sensors with such omniscient new technology? John Du of China Academy of Sciences gave the answer: "In the abstract, V2X technology itself is just a connection, and driverless driving is actually a concrete application."

Relationship between V2X and Autopilot

Physical sensors can be simply understood as mobile phone apps, while V2X plays the role of the underlying platform. Before the realization of this vehicle-road collaborative platform, the on-board sensor can only serve each mobile phone separately. In order to enhance the persuasiveness, on the V2X vehicle networking communication technology experience day, GM showed several safety applications to the media, including: vehicle blind area and lane change warning, intersection collision warning, emergency braking warning, front collision warning, vehicle speed guidance, front sharp turn warning, road wet warning and so on.

After actually getting on the bus and experiencing the universal V2X technology, Yiou Automobile can find that because vehicles need to accurately identify the traffic lights at intersections, know whether there are vehicles at the intersections ahead, and even predict the objects behind the corners, it puts forward higher requirements for the accuracy of the map return data. From a technical point of view, compared with assisted driving, the realization of V2X technology depends more on high-precision maps, which is the premise of vehicle positioning.

Obviously, these functions overlap with the driver assistance system to a great extent. Theoretically, based on dedicated communication technology and positioning technology, the effective communication range of V2V technology can reach 300 meters. Compared with sensor devices such as cameras or radars, vehicles equipped with V2V technology not only have a longer communication distance, but also can obtain long-distance road information through vehicles equipped with V2V technology, which is not easily affected or interfered by weather, light or cleanliness, terrain and obstacles.

It seems reasonable, but it is better than what Musk said: "The Tesla Autopilot version has been able to achieve fully automatic driving. At present, vehicles can change lanes and overtake on expressway, go up and down the slope, and soon they can realize the functions of identifying traffic lights and circling the island. "

They didn't mention V2X at all. In other words, can L3 self-driving vehicles without V2X technology also communicate with infrastructure such as signal lights? Strictly speaking, Musk has not clearly defined the concept of "completely autonomous driving", which cannot be directly understood as L3 level. What's more, the above commitments seem quite radical, but there is still no time node for the final landing.

To take 10,000 steps back, although John Du has always emphasized that practitioners should be open-minded, once Tesla really realizes full-automatic autonomous driving, it is almost equivalent to overthrowing the necessity of the existence of lidar and V2X, which practitioners have invested a lot of energy at this stage.

All use the same standards, and user experience is a competitive advantage.

In addition to Musk's "one bite, one pit" technology, the establishment of industry standard system is a bigger mountain on the road of large-scale application of V2X.

According to John Du, at present, China has completed the seven-layer communication protocol construction of V2X standard system, lacking the last step of security authentication system. After the standard system is basically improved, large-scale testing is needed at the application layer, which is expected to be combined with 5G early next year. The related work will be carried out under the guidance of the Ministry of Industry and Information Technology, and GM is also a member of the working group.

Basically, this set of standards is the first criterion for car companies including GM to implement V2X technology. To put it bluntly, the biggest role of the so-called communication standard in the whole industrial chain is to present the compliant information to the players and decide how to apply it. Let's take a mobile phone as an example. What the 4G network presents to operators is picture data information. At the beginning of the era of feature phones, operators thought of MMS. With the popularity of smart phones, the form of MMS has also been replaced by various applications such as WeChat.

In fact, the 4G network provides the same batch of data. The only difference is that WeChat gives users a better experience. Similarly, once the road data is uploaded to the vehicle, how the vehicle can provide users with early warning and other functions depends entirely on the manufacturer itself, which is also the biggest competitiveness of general V2X products in the eyes of John Du.

Compared with the competitive relationship, GM pays more attention to the shaping of user experience in product planning. "Different manufacturers have different understandings of traffic safety applications. V2X embodies the process of interconnection, which is very different from other autonomous driving technologies. For example, it is also an early warning function. Does it allow the owner to receive information through the screen or vibrate through the seat? Is it necessary to connect to the AEB system? This is the manufacturer's own design choice. "

From a technical point of view, V2X has successfully completed the demonstration and is expected to be rolled out on a large scale in two years. Different from the technical route that Musk can make money by himself, the foundation of realizing V2X lies in the cross-vendor cooperation of the whole network and the whole industry. As a century-old factory, GM is quite clear about this.

On the one hand, SuperCruise provides driving assistance to satisfy the appetite of the consumer market, and clearly tells the driver: "I can free your hands under certain conditions, but you are always responsible for the driving process of the vehicle." On the other hand, we never let go of L4-level unmanned fully-automatic driving, and constantly set up roadside units to receive subsequent communication standards.

If you really can't wait, GM still has one way to go: go straight to Robo-taxi, cut off the steering wheel and pedals, and spend $654.38 billion to mass-produce Cruise's brand-new self-driving car. It is planned to complete the San Francisco deployment on 2065.438+09.