"Smart Car" Meets "Smart Road"
Open the "Carrot Run" APP on your mobile phone, enter the starting point and destination, and click "Call Vehicle". A few minutes after the system accepted the order, a white self-driving car slowly stopped in front of the reporter.
Unlike the usual online car-hailing, there is no driver in the main driver's seat of this car, only a safety officer sits in the co-pilot seat.
Scan the code to unlock the car, click "Start Trip" on the display, and an autonomous driving journey begins.
Since the world's first high-level autonomous driving demonstration zone landed in Beijing Economic and Technological Development Zone, 300 high-level autonomous vehicles of various types have been regularly tested and commercialized within the area of 60 square kilometers, and the cumulative number of autonomous driving tests The mileage exceeds 4 million kilometers.
A few days ago, Beijing issued a notice of unmanned manned demonstration application, and the first batch of unmanned vehicles carried out demonstration applications, which means that the "unmanned behind the wheel" automatic driving service appeared in China for the first time, and Beijing became the first domestic passenger car service. A pilot city for unmanned vehicle operation.
What is it like to ride in a self-driving car?
How about safety and comfort?
In the future, is it possible for autonomous driving to replace human driving?
On May 23, the reporter came to the high-level autonomous driving demonstration area in Yizhuang, Beijing, for personal experience and on-site interviews.
"No driver behind the steering wheel" is the first attempt in China, and autonomous driving technology has moved from the laboratory to the real application market
In the Yizhuang High-level Autonomous Driving Demonstration Zone in Beijing, hundreds of Baidu self-driving travel service vehicles and other social vehicles are serving local residents every day from 7:00 am to 10:00 pm.
Recently, 20 "autonomous" autonomous vehicles from the "Carrot Run" platform, carrying cameras, lidars, sensors and other equipment, can complete starting, speeding, merging, and following on their own. , turning, avoiding obstacles, pulling over and other operations.
Sitting in a self-driving car, if you don't look at the driver's seat, the passenger will hardly feel that the robot is driving.
The steering wheel rotates automatically, stops slowly at the red light, controls the speed at the intersection and turns, and sounds the horn moderately. A series of "silky" operations make people feel novel.
The lidar on the roof and the cameras on the body are the "eyes" of the self-driving car.
There are three display screens in the car, two rear display screens and one central control screen. The screen of the display screen will not only simulate the driving route of the vehicle in real time, but also display the traffic in front of the vehicle such as pedestrians, trucks, and cars in real time. unit.
There are also multiple sets of cameras in the car, which will transmit the monitoring images back to the background control center in real time to ensure the safety of vehicle operation.
"Most users start their first experience with curiosity to see if no one is holding the steering wheel, stepping on the accelerator and braking, and many parents bring their children to experience this new thing." Baidu Intelligent Driving Business Group Vice President and Chief Safety Operations Officer Wei Dong said that in the first quarter of this year, the "Carrot Run" platform provided 196,000 rides. Through the analysis of user orders, it was found that the morning and evening peak orders were the largest, and the pick-up and drop-off points were concentrated. In subway stations, hospitals and shopping malls, "this shows that more and more people are beginning to use self-driving cars as a supplement to travel tools, and self-driving technology has moved from the laboratory to the real application market, which is the result we hope to see. ."
According to the reporter's understanding, in addition to Beijing, the "Carrot Run" platform has also opened autonomous driving travel services in Shanghai, Guangzhou, Shenzhen, Chongqing, Changsha and other cities.
According to the policy requirements, whether sitting in the main driver or the co-driver, the safety officer is still the standard configuration of the self-driving vehicle.
However, during the test ride of the reporter, the safety officer did not perform any operations except to remind passengers to scan the Beijing Health Treasure.
"The main responsibility of the safety officer is not to drive, but to observe the performance of the robot as an invigilator. Their hands are on their knees, but they are ready to take over at any time. If there is a very special situation, the safety officer will intervene by human driving. , to ensure the safety of passengers." Wei Dong said, "the main and co-drivers are unmanned" autonomous driving is trying to explore, with the maturity of technology and environment, the future will be completely unmanned.
The safety of autonomous driving is ensured through massive testing and redundant configuration, and the traffic efficiency of urban traffic is also improved
As an important turning point in the development of autonomous driving, "no one behind the wheel" is a manifestation of the maturity of the technology, and it is also an important node for commercialization.
In the future, as security officers gradually withdraw from the stage, can the personal safety of passengers be guaranteed?
What should I do if an emergency occurs?
"Humans need a driver's license to drive, and robots also need a 'driver's license'." Wei Dong said that every self-driving car must go through strict tests to ensure safety.
In addition to hundreds of simulation tests conducted by the R&D team, these vehicles must pass a professional assessment designed by a third-party agency hired by the government before they can obtain road qualification.
In addition, redundant configuration is also an important guarantee for vehicle safety, including sensor redundancy at the vehicle end, 5G cloud redundancy, lidar redundancy, and camera redundancy.
To put it simply, it is to provide the vehicle with "double insurance" or even "three insurances" through more configurations. If there is a problem in a certain link, the spare redundant configuration can be installed in time to ensure the normal driving of the vehicle.
In the case that the manual cannot control the vehicle in time, the car will also perform operations such as decelerating and safely pulling over to the side according to the degree of urgency, and then taking measures such as double flashing, honking, and automatically contacting the call center to ensure the safety of the people on board and the surrounding vehicles.
Statistics on road traffic accidents show that in 2020, a total of 244,674 road traffic accidents occurred nationwide, resulting in 61,703 deaths.
Among them, the five violations that cause major traffic accidents are: failing to give way as required, speeding, driving without a license, driving in the wrong direction and driving under the influence of alcohol.
And some of the problems that are prone to occur in human drivers do not exist for machines.
For example, the automatic driving system has an innate "rule awareness" and will not grab the lane or leave the car; without considering the charging, the automatic driving system can theoretically operate uninterrupted around the clock, and there is no fatigue driving. Problems, its movement patterns are basically all predictable and simulated; relying on the system's real-time perception ability, autonomous vehicles can also identify potential dangers of each traffic participant in advance, so as to better respond.
With the help of autonomous driving technology, people can travel to their destinations more easily in the future, and more importantly, it can greatly improve the efficiency of urban traffic and solve the traffic jam problem that plagues car owners.
Wei Dong analyzed the main "blocking points" in traffic traffic in three cities: First, the congestion caused by traffic accidents. Autopilot can identify the distance between vehicles very accurately, so it can avoid a lot of routine accidents such as scratches and rear-end collisions. .
The second is the blockage of the fork, which is often caused by the road grabbing and blocking between drivers. With the popularization of autonomous driving, the interaction between robots follows the same set of traffic rules, which can calculate the fastest way to pass and reduce the number of intersections. consume.
In addition, due to the limited parking spaces nearby, hospitals, shopping malls, scenic spots and other crowded places are prone to traffic jams, while self-driving cars can find and go to further parking spaces by themselves, and call the vehicle in advance when people need to use it. Reduce the hassle of looking for a parking space everywhere.
From a long-term perspective, after getting rid of the time limit for finding a parking space, the crowded spots can reduce ground parking spaces and introduce more three-dimensional parking lots, thereby reducing the consumption of urban land.
With policy support, infrastructure development, and 5G assistance, autonomous driving technology will be more widely used in transportation, logistics, agriculture and other fields
As the future trend of the development of the automobile industry, autonomous driving continues to "fly into the homes of ordinary people" from the laboratory. Strong policy support, good road conditions, diversified application scenarios, and the rapid development of communication technology under the background of 5G are indispensable.
Compared with European and American countries that paid more attention to bicycle intelligence in the early days, the development of autonomous vehicles in my country has always emphasized intelligent network connection and vehicle-road coordination. A series of strategies related to the development of intelligent networked vehicles, such as "Manufacturing 2025" and "Outline for Building a Powerful Transportation Country", have had an international impact.
As of November 2021, 38 provinces and cities across the country have issued detailed management rules, built 70 test demonstration areas, opened more than 5,200 kilometers of test roads, and issued more than 1,000 test licenses. Progress.
From traditional OEMs such as Hongqi, Changan, BAIC, and Geely, to communication companies such as Huawei, ZTE, and Datang, to Internet companies such as Baidu, Ali, and Tencent, to innovative companies such as Neolithic and Zhixing, major Enterprises are scrambling to deploy autonomous driving industry technology, and are actively trying to test and demonstrate operation areas in various places.
At the same time, the huge market scale has created a development space for the accelerated deployment of autonomous driving.
According to the "2021 China Mobile Travel Market Data Report", the online car-hailing market in 2021 will be about 319 billion yuan, an increase of 18.54% year-on-year; the user scale will be about 452 million, a year-on-year increase of 23.83%.
In the future, the market scale of hundreds of billions of yuan and the user group of hundreds of millions will become an important scenario for the commercialization of autonomous driving.
The application of autonomous driving technology is just the beginning.
From unmanned trucks on duty in mines and ports, to unmanned vehicles on the front lines of anti-epidemic transportation of anti-epidemic materials and daily necessities, it is deeply integrating with daily life.
In the paddy field of Hongyi Grain Planting Professional Cooperative in Xinzhan Town, Zhaoyuan County, Heilongjiang Province, an unmanned rice transplanter is automatically moving forward with a full load of seedlings.
Although it is unmanned, the rice transplanter "walks" very smoothly, leaving neat rows of seedlings wherever it goes. Calculated at a cost of 260 yuan per person per day, about 20 days of transplanting cycle can save more than 5,000 yuan.
In the office buildings of Suzhou High-speed Railway New City, a group of special "courier boys" shuttles through them every day. The sender and recipient only need to fill in the relevant information through the mobile APP, and then they can book an unmanned delivery vehicle to "run errands".
They have a maximum load of 400kg and can adapt to various scenarios such as indoors, outdoors, underground garages, and overhead floors.
On the urban streets of Nansha District, Guangzhou City, Guangdong Province, driverless sanitation vehicles can carry out various urban sanitation operations such as road cleaning, watering and dust reduction, and can complete daily routines such as automatic charging, water filling, sewage discharge, starting, parking, etc. without human intervention. maintenance work.
In order to assist the city to do a good job in epidemic prevention, the vehicle has added a spraying and disinfecting function to achieve zero-contact disinfecting in the whole process.
"The future of autonomous driving has unlimited possibilities. In the new round of global industrial competition, China has a unique competitive advantage. It has both the background of new infrastructure and a mature 5G network. I believe that whether it is the automotive industry or other industries, no one will Driving skills will be promising." Wei Dong believes.
"Smart car + smart road", vehicle-road coordination makes "green light freedom" possible
The experience of "green light all the way" can make car owners feel full of happiness on the road.
Today, "green light freedom" is no longer only based on luck, but is achieved through vehicle-road coordination.
The so-called vehicle-road coordination refers to the connection between the car end and the road end using the network and mutual communication through cloud computing.
For example, when the vehicle is driving on a road with a signal light, when the red light will be on, the road equipment will use the network to transmit this information to the vehicle. Through Baidu map navigation prompting the signal light to count down, the vehicle can control the speed in advance and reduce the Number of stops.
Even when the road equipment detects that a pedestrian suddenly runs a red light, it can notify the vehicle in advance to avoid accidents.
From the construction of digital infrastructure such as smart light poles and smart intersections, to the two-way empowerment of roadside data, cloud data and vehicle end, to the implementation of smart signal control, smart off-site law enforcement and other applications, a series of innovations are being piloted across the country The city landed.
As of May this year, Beijing Economic and Technological Development Zone has built 332 intelligent intersections, covering 60 square kilometers in the Economic and Technological Development Zone.
At present, citizens can enjoy the "green light all the way" travel experience on 13 arterial roads including Xihuan Road, Ronghua Road, Rongjing Street, Yongchang Road, Tongji Road and Hongda Road in the main urban area of the Economic Development Zone.
In Baoding, Hebei, Baidu has intelligently upgraded hundreds of intersections, enabling traffic lights to have the ability to "see six roads and listen to all directions", obtain the traffic flow status at intersections, calculate and analyze the law of traffic evolution, and thus real-time dynamic Adjust the timing strategy of traffic lights.
The peak traffic congestion index in Baoding urban area has dropped by 4.6%, and the average speed has increased by 11.6%.
"Smart car + smart road" can maximize the effect of vehicle-road synergy.
According to the intelligent level of autonomous driving technology, the International Society of Automotive Engineers (SAE) divides the level of autonomous driving into 6 levels of L0-L5, that is, from fully manual driving to fully autonomous driving.
At present, how "smart" are our self-driving cars?
"The watershed of autonomous driving lies in L3." Wei Dong introduced to reporters that the intelligent core of L3 and below is people, and people always maintain the initiative of driving. For example, the cars with automatic parking and autonomous cruise in my country's market are generally L2. level.
The L4 and L5 levels are autonomous driving of machines. For example, the self-driving car of the "Carrot Run" platform is L4, which still requires a certain amount of manual intervention, while L5 hardly requires any manual intervention.
How long will it take to achieve L5-level intelligent driving in China?
Wei Dong believes that the realization of L5-level autonomous driving should respect the rigor of science, and the demonstration should be cautious and not in a hurry.
To achieve L5-level autonomous driving, there must be a good ecology, which includes both policy elements and infrastructure elements, including technological maturity and interconnectivity.
According to the current path of technological evolution and the maturity of the supporting conditions, it is judged that there will be opportunities to promote the scale of L4-level autonomous driving from 2024 to 2025, and all new models manufactured by 2040 will have autonomous driving capabilities.