Autonomous driving has "taken orders" the next stop for large-scale commercialization

As a high-level autonomous driving demonstration area with networked cloud control, Beijing Economic Development Zone took the lead in launching the country's first pilot service for the commercialization of autonomous vehicles within 60 square kilometers. This marks that domestic automatic driving has entered a new stage of commercial pilot exploration from test demonstration, and automatic driving has begun to integrate into public life.

"The driving is smooth and the price is good, but I didn't expect that such high-tech really entered our lives and brought convenience." On November 25, 2021, Ms. Yuan, a resident of the Beijing Economic and Technological Development Zone (hereinafter referred to as the Beijing Economic and Technological Development Zone), posted a screenshot of a 1.34 yuan order for a self-driving taxi in the circle of friends and social networking sites, which was frequently liked. This is the first commercial landing order after Beijing opened the first pilot commercialization of autonomous mobility services in China.

Just before the completion of this order, Beijing Economic Development Zone, as the world's first networked cloud-controlled high-level autonomous driving demonstration area, took the lead in carrying out the country's first pilot service for the commercialization of autonomous vehicles within 60 square kilometers. The first batch of licenses for the commercialization pilot of autonomous driving travel services was issued to Baidu Apollo and Xiaoma Zhixing, which marked a new stage of domestic autonomous driving from test demonstration to commercial pilot exploration, and automatic driving began to integrate into public life.

Play a role in the commercial operation of autonomous driving

With the landing of the first commercial order for autonomous driving, paying for intelligent and connected passenger cars (Robotaxi) has become the daily travel choice of residents in Beijing Economic Development Zone. "The opening of Robotaxi payment means that the industry has opened the verification of the commercial model of autonomous driving, and driverless technology will also enter the travel service market in a more product-oriented form." The relevant person in charge of Xiaoma Zhixing said, "In the more than 1 month of commercialization pilot, Robotaxi has had hundreds of paid orders every day. ”

The commercialization of Robotaxi is a recognized difficulty in the industry, and the full-scenario L4 automatic driving is a peak that is difficult to overcome, that is, to achieve highly automatic driving, the whole process does not require any operation by the driver, generally relying on the road information data support that can be updated in real time, in the real scenario of travel, complete the automatic pick-up and return, automatic formation cruise, automatic obstacle avoidance, etc.

So, how does artificial intelligence play a role in the commercial operation of autonomous driving? "At present, software systems using technologies such as collective artificial intelligence on Robotaxi can realize system perception without blind spots." The relevant person in charge of Xiaoma Zhixing said.

According to the technical staff of Xiaoma Zhixing, the traffic light recognition camera can help Robotaxi identify the status of the signal light more accurately, and the lateral millimeter-wave radar is used to monitor the situation of the left straight car when the vehicle turns right... On the Robotaxi of Pony Zhixing, through 4 lidar, 4 millimeter wave radar, 7 cameras, and two GNSS antenna modules, it can achieve 360-degree blind spot perception, providing a more optimized solution for long-tail scenes and complex scenes. The vehicle transmits the perceived valid data back to the background during the driving process, and the artificial intelligence and deep learning network will mine the road test data and carry out automatic annotation and learning training, so as to drive the rapid iteration of the perception algorithm and prediction module.

"To realize the commercialization of Robotaxi, we not only have the technical route support of bicycle intelligence, but also the escort of the technical route of vehicle-road collaboration." Baidu Apollo, another company that has been licensed to pilot the commercialization of autonomous mobility services, has both a technical route that can realize the intelligence of bicycles through artificial intelligence technology, and a collaborative technical route of vehicles and roads that are closely connected with urban transportation. In terms of bicycle intelligence, Baidu Apollo also launched the only L4-level visual perception solution for urban roads in China. In the field of vehicle-road collaboration, Baidu Apollo proposed Apollo Air, a vehicle-road collaboration technology that can realize the closed-loop of L4 autonomous driving of the open road continuous road network solely by relying on roadside perception. Not only that, Baidu Apollo will also apply artificial intelligence to the "5G cloud", when there is a temporary road change or traffic control, etc., after receiving a request for help from an autonomous vehicle, the "5G cloud driver" can take over the driverless car and quickly solve the problems in the vehicle's travel.

Realize urban high-speed transportation and end-of-line distribution

At noon on January 4, in the office of Chaolin Building in Beijing's Jingkai District, Mr. Tian opened the Neolithic Special Delivery Mini Program, and the four kinds of meals provided by the Neolithic unmanned vending vehicle downstairs were displayed in front of his eyes. "It's too late to eat at noon, just look at the small program to see the unmanned car."

What are being sold, go downstairs to scan the code to pay for it, which is particularly convenient. He said. Wave your hand, scan the code to pay for meals, every day during breakfast and lunch, the neolithic unmanned vending vehicles are the "red people" of the park, which solves the problem of eating for the staff of the park. In fact, under this office building, there are unmanned delivery vehicles coming and going every day, there are couriers, takeaways...

With the rapid development of artificial intelligence and autonomous driving technology with human-like cognitive decision-making ability, the process of autonomous driving commercial landing has gradually accelerated, which is covering the last 3 kilometers of urban end distribution, embracing the huge market opportunity of unmanned distribution.

Only to see the self-driving truck drive to the entrance of the highway, slow down the speed, follow other vehicles into the highway; as the traffic becomes larger and smaller, adjust the speed in time; encounter the curve, slowly turn and speed up... On the last day of 2021, the self-driving trucks of Mainline Technology and Xiaoma Zhika have been on the road, opening the "high-speed" era of automatic driving and taking an important step towards commercial operation.

"The ability to make self-driving trucks drive so smoothly on the highway is due to our self-developed artificial intelligence core technology, which makes the truck seem to have a brain, and can accurately identify the surrounding environment and make corresponding driving judgments." Wang Chao, dean of the Mainline Technology Prospect Institute, said, "Doing automatic driving requires a lot of testing, constantly excavating long-tail scenarios, and realizing data closed loops, all of which are in order to allow autonomous vehicles to drive safely in more scenarios." ”

Based on the self-developed core technology of "virtual driver", the Xiaoma Zhika autonomous truck system adopts a multi-sensor fusion scheme to achieve 360-degree perception and centimeter-level positioning, of which the long-distance camera covers a range of 1,000 meters. Self-driving trucks can fully detect the road conditions in front of high speed through artificial intelligence technology, and the rear camera can help vehicles change lanes more safely in high-speed scenarios, at the same time, through the optimal path and intelligent combination of throttle, brake and other control schemes, autonomous trucks can achieve safe and stable driving, improve vehicle operation efficiency.

Wang Chao said: "Open highway testing can help enterprises accelerate the iterative upgrading of algorithms and products, collect a large amount of data, promote the rapid development of the industry, and usher in the dawn of commercialization of autonomous trucks." ”

Obviously, with the blessing of artificial intelligence, unmanned driving can be slow and fast, and it is realizing the high-speed transportation and terminal distribution required by the city.

Explore viable business models for autonomous driving

Safety is the foundation of autonomous driving. Undoubtedly, under the premise of ensuring safety, we can explore a feasible business model for autonomous driving.

According to reports, the Beijing High-level Autonomous Driving Demonstration Zone gives a "Chinese plan" for science and technology to empower the city, that is, to build five major systems of "smart roads, smart cars, real-time clouds, reliable networks and accurate maps".

In order to ensure the safe operation of automatic driving, the Beijing High-level Autonomous Driving Demonstration Zone, which is currently launched in the Beijing Economic Development Zone, has completed the construction task of Phase 1.0. That is, through artificial intelligence technology, the real-time integration of vehicle and road perception data is realized at the end of the car, so that the lane-level accuracy rate is greater than 99%; the standard intersection construction scheme of intelligent networked vehicles is mainly realized at the end of the road, which increases the traffic efficiency by more than 30%; in the cloud, realizes the digital holographic intersection and completes the definition of 17 types of traffic events such as faulty vehicles; at the network end, with the help of highly reliable, low-latency EUHT network, real-time transmission of traffic management information, real-time transmission of vehicle-road collaborative information, remote monitoring of parallel driving, Vehicle remote driving and other functions; at the end of the map, a high-precision dynamic map platform is built to respond to the industrialization needs of enterprises and explore commercialization.

"In the next stage, the demonstration area will continue to summarize the previous construction experience, focus on the implementation of the five special projects of 'road expansion, strong cloud, networking, inter-vehicle connection, and map landing', achieve full coverage of digital intelligent intersection infrastructure within 60 square kilometers of the Beijing Economic Development Zone, and promote the innovative application of technologies such as collaborative innovation and intelligent traffic management." Kong Lei, deputy director of the Management Committee of Beijing Economic Development Zone and director of the Beijing High-level Autonomous Driving Demonstration Zone Office, said.

It is understood that in terms of the management policy system, the demonstration area will focus on the specific scenarios in the application of intelligent network passenger buses, unmanned shuttle cars, unmanned sweepers and other specific scenarios in the application on the basis of the policies that have been introduced, as well as common issues such as the division of responsibility for intelligent networked vehicle accidents and data security management.

Kong Lei pointed out: "We will continue to carry out scenario innovation, improve the cloud control application ecosystem, provide city-level services for intelligent and connected vehicles, support innovative applications in multiple industries, and create a 'Beijing solution' for city-level applications of intelligent and connected vehicles." ”

Source: Science and Technology Daily