Although the future of “L5 fully automatic driving” seems bright, the road is still tortuous.
Editor’s note: This article is from WeChat public account “AIDaily“, author Binbin, Edit 仟 billion, small a.
On August 2nd, Baidu and FAW Hongqi jointly developed the first domestically produced L4-class self-driving taxi Robotaxi-Hongqi E·jie debut in Changsha. This year, Changsha City will open the road manned test for L4 self-driving vehicles. This news quickly swept the major websites, and the most critical “L4 level autopilot” caused great concern.
About the classification of autonomous driving, it is divided into SAE (American Society of Automotive Engineers) and NHSTA (National Highway Safety Administration) two levels of standards, of which SAE divides the automatic driving level into 6 levels. From L0-L5, NHSTA is divided into 5 levels from L1-L5.
The difference between the above two sets of standards is mainly for the definition of L4 and L5. In the classification of NHTSA, L4 is already fully automatic driving, while SAE International’s L5 standard stipulates that vehicles must be able to cope with all situations. Including drastic driving, you can also pass smoothly.
When the L3 level of autonomous driving technology just started, “L4 automatic driving” was widely sought after. Many users mistakenly thought that L4 could achieve full automatic driving, but “L4” can make the vehicle “not at all.” Human intervention is required to complete the established driving tasks on its own, but it must be based on the premise of “limited roads and environment”, instead of carrying the passengers to truly open the “flying life” of the river.
“In the limited roads and the environment” not only requires the vehicle itself, but also the road conditions. At present, there are not many roads that can fully satisfy the L4 self-driving. Most of them are used for auto-driving car testing. This is also the main reason why China’s relevant laws do not allow auto-driving cars to go on the road, losing the road’s “hard core.” “Support, autonomous driving is the standard “road killer.”
“L&D” jointly developed by Baidu and FAWThe 4th-level self-driving car is the L4 in the SAE system, namely: “Super-high-automatic driving”, unlike “L5-level fully automatic driving”, “L4” cannot cope with all road conditions, only on specific roads or Driving in a specific environment. Although the conclusion is inevitably disappointing, the results have indeed taken an important step toward full automatic driving of L5.
What is the difficulty of “automatic”
The development of autonomous driving technology is the biggest difficulty in L5-class automatic driving. More sensitive and efficient cameras, laser radar, sensors, touch panels, and on-board processors are all indispensable factors for achieving L5. More importantly, the pursuit of better development in autonomous driving ultimately comes down to social synergy. It is not that technology companies can rely solely on hardware and software technology.
At this stage, the autopilot technology used by each family has different processing under logic execution and emergency handling. Even when the same system faces the same danger again, it will react completely differently. It is inevitable that the vehicle’s own terminal power is insufficient or the efficiency is difficult to cope with complex traffic data, causing misjudgment and even causing accidents.
The bottleneck in the development of the autopilot industry is visible to the naked eye. The same industry standards have been growing. Only autopilot car systems can achieve the same IOS and Android systems as mobile phones. Good development, not every one of us now has its own way.
Industry believes that to accelerate the development of the autonomous driving industry, the simultaneous road requirements must be rapidly upgraded to achieve better vehicle information interaction, so as to more effectively improve the stability and safety of autonomous driving. Therefore, clear and clear road signs and ground lines are essential for the development of autonomous driving.
Industry believes that autopilot alone cannot completely avoid security risks in a short period of time due to sensors that are limited by the vehicle. Only by realizing the synergy of vehicles and roads can an efficient and safe intelligent transit vision be realized under complicated road conditions.
For example, the road itself can monitor the position of all driving vehicles and pedestrians and synchronize them to the car. Then, after the interaction of the system, the safety of the driver and pedestrian can be better ensured, which is greatly shortened. Travel time, even in the future