China News Service, August 18 (Zhongxin Finance and Economics Ge Cheng) Recently, a traffic accident related to the assisted driving system of Xiaopeng Automobile has made the safety of the assisted driving system hotly discussed again.
In fact, not only Xiaopeng Motors, but also new car-making forces such as Ideal and Weilai have been on the hot search for accidents related to assisted driving systems.
Zhongxin Finance noticed that in many traffic accidents caused by the use of assisted driving systems, many vehicles encountered stationary objects while driving, and the brakes were insufficient or did not brake.
Focus 1: The assisted driving system cannot recognize stationary objects?
A stationary object that can be easily recognized by the human eye is not "seen" by the assisted driving system all over the sensor radar?
"The adaptive cruise system relies on millimeter-wave radar and visual sensors for target recognition." The person in charge of technology development of an autonomous driving company told Zhongxin Finance, "The system is not unable to identify stationary objects, but cannot distinguish stationary objects. objects and other distractions."
"The current millimeter-wave radar technology cannot distinguish stationary vehicles from roadside pillars and signs." The person in charge made a further interpretation, "If such faulty vehicles are not filtered, all stationary objects will be regarded as Obstacles, which often lead to erroneous braking.”
"From a technical point of view, it is not bad that the current assisted driving system can prevent accidental braking. In the process of development, we mainly try to eliminate as many accidental braking as possible." Zhu Xichan, a professor at the School of Automotive Engineering of Tongji University, was interviewed. Tell Zhongxin Finance, "If you can't brake when you shouldn't brake, there will be a lot of brake leakage."
"The assisted driving system's recognition of static objects on the road is just the opposite of that of humans. It is easier for human drivers to recognize static objects." Zhu Xichan further explained that because the camera has an optical flow algorithm, millimeter-wave radar has many Puller effect.
Whether it is a millimeter-wave radar or a camera, it can recognize moving objects.
Focus 2: Who misled assisted driving?
Although the assisted driving system is currently facing difficult identification problems, many consumers and users have high expectations for assisted driving, thinking that as long as the hardware is sufficient, they can drive automatically.
Therefore, some videos about "unmanned driving" on social media are not uncommon, which are also related to the inductive propaganda of some car companies before that the assisted driving system is an intelligent automatic driving system.
In fact, "assisted driving" does not mean "autonomous driving", and it is not safe to use the assisted driving system for "unmanned driving".
The assisted driving system can indeed make our "tight" nerves "relax" a little while driving, but it is difficult to be distracted, and we still need to be aware of accidents at any time.
"At present, there is no domestic company that can sell self-driving cars." Zhu Xichan told Zhongxin Finance that many accidents stem from driver misuse.
"Why do you have to emphasize assisted driving? It's because you can't figure it out."
"When the assisted driving system gives the driving power to the system, the driver cannot be distracted. Even though the driving power is handed over to the car, you still have to concentrate on driving." Zhu Xichan said that whether it is encountering someone who fell on the road The car is still parked on the road, the system does not brake, the driver must brake.
When driving, you have to keep your hands on the steering wheel, and if you find that he is wrong, he will take over immediately.
The current level can only do so much.
Focus 3: Are we only three years away from "autonomous driving"?
How far are we from autonomous driving systems that allow drivers to take their hands off the steering wheel?
At present, there is a view in the industry that the widely adopted intelligent driving technology for bicycles has encountered a bottleneck, and autonomous vehicles need new technology routes such as vehicle-road coordination.
On the premise that vehicles and roads are equipped with sensors, the autonomous driving system will be able to obtain more comprehensive road traffic data, and the interconnection between vehicles can also correct the data errors provided by a single sensor to the greatest extent.
Li Jun, an academician of the Chinese Academy of Engineering, said, "Under the technical premise of vehicle-road coordination, all-weather and all-round autonomous driving will be truly possible."
Terminal perception and 5G short-term communication can indeed improve the autonomous driving capability of the vehicle, and even reach the unmanned standard.
But in Zhu Xichan's view, it's still too early.
"Vehicle-road coordination is only tested and demonstrated in the demonstration area. It needs the support of infrastructure, and it does not yet have the conditions for large-scale promotion."
"In 2021, my country's L2-level (combined driving assistance) assisted driving passenger car market penetration rate will reach 23.5%, and it will increase to 30% in the first half of this year (2022)." At the press conference of the Connected Vehicle Conference”, Wang Weiming, director of the First Division of Equipment Industry of the Ministry of Industry and Information Technology, said that at present, the test mileage of open roads across the country has exceeded 5,000 kilometers, and high-level intelligent connected vehicles have taken the lead in specific and limited scenarios. The test and demonstration of large-scale manned cargo was carried out.
"Typical L3 (conditional autonomous driving) is still possible in recent years." Zhu Xichan expects, "It is estimated that in about three years, there should be such a car to drive." (End)