China is leading the world in unmanned driving, and one of the issues that everyone is concerned about is how safe it is. In fact, we should never have thought that there would be such a driverless car before, but we in China did it, and we are far ahead of the whole world. Only we in China have realized the driverless technology. In fact, at present, the development trend of the global driverless car industry is relatively good, but there are few areas for mass production. Maybe this technology is immature or everyone is afraid to put it into mass production, because after all, driverless driving is a very dangerous technology, and if it is not done well, the consequences will be very serious.
This driverless technology was developed together with new energy vehicle related technologies of 5G communication technology. In fact, the research and development of this driverless car has entered the stage of trial operation and debugging, and most domestic R&D companies are in the experimental stage. With the continuous maturity of driverless technology and the adjustment of government policies, it is believed that such driverless cars will soon operate in industrial development and then be used in commercial fields. It will eventually develop into civilian use, which means that all of us may use such unmanned vehicles in the future. I believe this day will come soon, because we always believe that national researchers will make this day come soon.
It is not an easy task to realize the technology of driverless cars. Maybe they have failed many times before and rehearsed many times. In fact, the main system of the core technology of driverless cars is divided into three levels: perception of nuclides and execution. If it is to be implemented, it will be very difficult. First of all, it must be able to perceive the surrounding environment like human eyes and ears, and it must be able to collect and process with the surrounding environment. In other words, such driverless technology should be able to cope with all kinds of emergencies like people.
By 2025, China will popularize unmanned driving? China 2025 plans to be unmanned.
; ? Recently, Hou, Deputy Secretary-General of China Automotive Engineering Society and Executive Dean of International Automotive Engineering Science and Technology Innovation Strategy Research Institute, said during the Shanghai Auto Show that China is expected to achieve 70% mass production of L3-class autopilot and L4-class autopilot in 2025 and L5-class autopilot in 2030.
Development history of autonomous driving
At present, the classification of autonomous driving is mainly based on the standards of SAE American Society of Automotive Engineers, and the autonomous driving technology is divided into six grades from LOLO to L5. Of course, other countries and institutions have also introduced relevant automatic driving level classification, but they are basically similar to SAE automatic driving classification standards. For example, the Ministry of Industry and Information Technology of China announced the recommended national standards for automatic driving classification on March 9, 2020.
To put it simply, the LO class, as the name implies, depends entirely on manual driving. Because this level of autonomous driving can be traced back to the birth of cars, there are often not many topics when discussing autonomous driving technology.
L 1 autopilot is basically constant speed cruise. Constant speed cruise can not only stabilize the speed, but also adjust the speed through the multi-function steering wheel. Although cruise control has no intelligence at all, functionally speaking, it is the embryonic form of the development of autonomous driving technology, and the hidden dangers brought by autonomous driving technology also begin with cruise control.
By the way, although cruise control can alleviate the fatigue of drivers during long-distance driving to a certain extent, cruise control can also relax drivers, and some drivers even leave their right feet off the brake pedal after cruise control. You know, if you relax your vigilance at high speed without the assistance of any detection equipment, it means that death is approaching.
L2 autopilot technology is to add detection equipment including millimeter-wave radar or camera on the basis of constant-speed cruise, which can realize the detection of road conditions and automatically complete the acceleration and deceleration control of vehicles. This is ACC adaptive cruise.
Interestingly, autopilot has developed rapidly in recent years, and L3-level autopilot related policies have not yet landed. Many car companies have also built an L2.5-level autonomous driving when promoting their own models. The so-called L2.5 autopilot technology is actually above the passing line of L2 autopilot, but below the L3 autopilot standard.
There are two main reasons for this situation: from the commercial point of view, with the gradual popularization of L2-level autopilot technology, the autopilot ability in L2-level autopilot category is uneven; In addition, the policy on L3 autonomous driving technology has been delayed; Therefore, the simple "L2" can no longer meet the publicity needs of automobile manufacturers with relatively high degree of autonomous driving.
For example, only vehicles equipped with non-full-speed ACC adaptive cruise can promote their L2-level autopilot technology; Vehicles equipped with full-speed ACC adaptive cruise+lane centering assist function, which can realize one-lane lateral+vertical automatic driving assist control, can only boast that they have reached L2 level of automatic driving technology. In order not to conflict with the policy and show the advantages of autonomous vehicles in autonomous driving, witty automobile manufacturers put forward L2.5-level autonomous driving between L2 and L3.
Why is the evolution between L2 and L3 so difficult?
When starting the L2 autopilot, the driver must hold the steering wheel, look forward and be ready to deal with emergencies at any time. Therefore, L2 autopilot is more like an intelligent version of cruise control, which is far from the real autopilot technology. In addition, literally, the so-called "L2 autopilot" actually hides a very important word-assistance.
"Auxiliary", a keyword that should have been displayed in bold, has misled too many consumers because of the propaganda methods of some car companies, which has cost many consumers their lives.
The biggest difference between L3 autopilot and L2 autopilot is that the definition of "assistance" is cancelled, which can really liberate the driver's hands. After the automatic driving is turned on, the system will completely take over the driving in limited scenes, and the driver's hands and feet are useless for the acceleration, deceleration, lane change, turning and other operations that need to be carried out during the driving process.
Because L2-level autopilot technology is only an auxiliary stage, at present, most car accidents caused by over-reliance on autopilot technology will basically be borne by the party using autopilot. If car companies, auto-driving technology suppliers and even the whole industry have affirmed the safety of L3-class auto-driving in a specific road environment, but can't guarantee its absolute safety, then the definition of L3 is meaningless.
This is a century-long problem that L3 autopilot technology is currently facing, and it is also the main reason why L3 autopilot related policies have been delayed. However, judging from Dean Hou's remarks during the Shanghai Auto Show, China has made a breakthrough in L3 autopilot.
What's the difference between autonomous driving and driverless driving?
Autonomous driving and unmanned driving are different cognitive subjects. If people decide driving behavior, it is automatic driving. Unmanned driving is a higher level than automatic driving, that is, the driving work is completely handed over to the machine, which is also called automatic driving.
Self-driving car, also known as driverless car, computer-driven car, or wheeled mobile robot, is a kind of intelligent car that realizes unmanned driving through computer system. It has a history of several decades in the 20th century. At the beginning of the 20th century, it showed a trend towards practicality.
Self-driving cars rely on the cooperation of artificial intelligence, visual computing, radar, monitoring equipment and global positioning system, so that computers can automatically and safely operate motor vehicles without any human initiative.
Driverless car is a kind of intelligent car, also known as wheeled mobile robot, which mainly relies on the intelligent driver with computer system in the car to achieve the purpose of unmanned driving.
According to Thomson Reuters's latest report on intellectual property and technology, from 20 10 to 20 15, there are more than 22,000 invention patents related to automobile driverless technology, and in this process, some enterprises have emerged and become industry leaders in this field.
2065438+September 2009, the first batch of L4-class self-driving passenger cars jointly built by Baidu and FAW-Hongqi EV was mass-produced in China, and five Beijing self-driving road test licenses were obtained.
On September 22nd, the National Intelligent Networked Automobile Test Demonstration Zone was officially unveiled, and companies such as Baidu, Liang Hai Technology and Deep Blue Technology obtained the world's first commercial license for self-driving cars. 20 19 On September 26th, Baidu announced in Changsha that Robotaxi, a self-driving taxi fleet, was officially put into trial operation.
Is a car with driverless technology safe and reliable?
20 16? Year? 5? At the end of the month, a European Tesla owner rear-ended a large truck on the highway. In addition, in the United States, a Tesla crashed into a parked trailer, and one was in? 5? Hit a moving vehicle on Interstate 10. All these accidents happened in? Autopilot? 20 16 is on? Year? 5? Month? 7? From Ohio, USA? 40? An old man drives a car? Autopilot? A model with automatic driving mode turned on? s？ When he reached an intersection, he died after an accidental collision with a large trailer. these
At present, I think driverless technology is unreliable, but the future development of driverless driving must be the general trend. Driverless cars are a major trend in the future development of science and technology, and this trend will become more and more obvious. At present, there are still many shortcomings in control technology, sensor technology and image recognition technology, and many improvements can be made. Whether it is Tesla conventional radar, ultrasonic radar and monocular camera, there are also powerful chip computing capabilities and the use of high-precision maps. It is also a lidar launched by Huawei. For complex road conditions and changeable weather, these sensors are not 100% accurate.
First, the detection range of on-board sensors is limited.
The finiteness mentioned here refers not to distance, but to height. As far as Tesla is concerned, his detectors are basically installed at the bottom of the vehicle. Compared with the higher chassis of a large truck, the suddenly rushed out truck body may not enter the detection range of sensors and radars, thus causing a collision.
Second, the error of the algorithm
Although unmanned driving has a large number of simulated road driving and powerful AI intelligent algorithms, through the analysis of relevant people, several previous accidents can be preliminarily judged. In some cases,
But I always believe that unmanned driving is the future trend, and accidents will happen when people drive vehicles, some of which are caused by drunk driving or fatigue driving. These can actually be avoided, which can also reduce unnecessary traffic accidents when the driverless technology is mature in the future.
What is a driverless car?
Driverless car is a kind of smart car, which can also be called wheeled mobile robot. It mainly depends on the intelligent driver in the car, mainly on the computer system, and realizes unmanned driving.
Driverless cars integrate many technologies such as automatic control, architecture, artificial intelligence and visual computing. It is the product of highly developed computer science, pattern recognition and intelligent control technology, and it is also an important symbol to measure a country's scientific research strength and industrial level, and has broad application prospects in the fields of national defense and national economy.
Since the 1970s, the United States, Britain, Germany and other developed countries have begun to study driverless cars, and have made breakthroughs in feasibility and practicality. China began to study driverless cars in the 1980s, and the University of National Defense Technology successfully developed the first truly driverless car in China at 1992.
In 2005, the first urban driverless car was successfully developed in Shanghai Jiaotong University. It is expected that the car will be the first demonstration operation in Shanghai Century Park within two years, and will show its talents at the 20 10 World Expo. At that time, visitors only need to press a button at the gate of the park, and a four-seat convertible car without a driver will come from a distance, stop slowly, and then take the passengers to the scenic spots they want to go.