Autopilot is different. How many ways do they play？
[Aika Automotive Technology Original]
Speaking of autonomous driving, it is no longer a new concept. Since childhood, self-driving cars have appeared frequently on the big screen, and now the self-driving technology has gradually matured and stepped out of the screen, all of which are inseparable from the rapid development of science and technology. However, today, with so many manufacturers developing autonomous driving technology, do you know how many ways autonomous driving can be realized? Next, let’s explore one or two.
At the state-of-the-art level,Automatic driving can be roughly divided into two technical routes.：
The first one is based on the existing radar and sensors of the vehicle, taking ADAS (Advanced Driver Assistance Systems) as the starting point, and conducting the research on automatic driving.. This technology tends to allow vehicles to perform real-time calculations, and through the cooperation of on-board facilities such as cameras and radars, vehicles can achieve the purpose of automatic driving.
The second one is based on V2V(vehicle to vehicle means information interaction between vehicles) and V2X(Vehicle to X means information interaction between vehicles and the outside world), which allows vehicles to establish information communication with the surrounding environment through mobile internet.So as to know the real-time traffic conditions and realize predictable automatic driving.
Nowadays,Both of these technologies have made great progress, but there is still a considerable gap from the real automatic driving.
1. Automatic driving based on ADAS:
Why use ADAS as the basis of autonomous driving? The reason is simple,Because ADAS already has a mass production foundation.The gradual penetration of various driver assistance systems into low-end models is the best proof.
At present, although the global market penetration rate of ADAS is not high, only about 5%, with the increase of automobile safety and intelligent demand, the market penetration rate of ADAS will be greatly improved in the future. According to the prediction of consulting firm PRNewswire, by 2020,The global ADAS penetration rate will reach 25%, and the global new car ADAS loading rate is expected to exceed 50%.Isn’t it enough that such a huge penetration base is used as the basis of autonomous driving?
Moreover, ADAS is quite mature in hardware.Bosch, NXP, Continental and other parts suppliers have already supplied tens of millions of sensors and actuators to car companies every year.Such a huge supply is enough to show that ADAS is quite reliable in hardware. In particular, NXP has recently released a powerful autopilot platform using the latest BlueBox engine.
But can you achieve automatic driving only by relying on ADAS? Of course not. Take Tesla as an example. Tesla is ahead in the field of ADAS, but it can only reach the level of second-class automatic driving. All this is due to the lack of effective support from high-precision maps.
Nowadays,High-precision maps have become a necessary and insufficient condition for autonomous driving.. According to the division of map manufacturers, high-precision maps can be roughly divided into two categories.The first is ADAS-level high-precision map.Its accuracy can reach the meter level;The second is the HAD-level high-precision map.Its accuracy can reach the centimeter level.
Because the two maps have different purposes,The ADAS level only includes basic information such as road shape and slope.And then used to cooperate with the driving assistance system of the vehicle itself to reduce driving pressure.The information collected by HAD level is more accurate, and it will record the lane type, width and surrounding environment of the road in detail.In this way, vehicles will have more detailed reference information, so on the basis of ADAS, high-precision maps have become an indispensable part.
However, due to the complexity of urban roads, it is quite difficult to collect high-precision maps in urban internal roads.It will take time to mature.
Second, autonomous driving led by the Internet of Vehicles:
After talking about the automatic driving based on ADAS, let’s talk about the automatic driving led by the Internet of Vehicles.
Simply put,The technical concept of V2V and V2X is to install a communication device on the car, and then use traffic lights, road signs, road vehicles and pedestrians as signal emission sources for real-time communication.. In this way, even if the car itself does not have the detection ability, it can know the traffic environment around it by constantly asking about the surrounding vehicles and traffic facilities, so as to realize safe automatic driving.
Compared with ADAS, V2V and V2X technologies have the greatest advantage that they can make the vehicle have foresight, instead of reacting in front of it.. Take the traffic light as an example. ADAS recognizes the image based on the camera, and judges whether the traffic light is red or green through the algorithm, while V2X can let the vehicle receive the signal 100 meters away, thus knowing the status of the traffic light, which makes the automatic driving easier.
However, this technology still needs high-precision maps as the foundation.Its biggest difficulty is that it is a very huge thing to install signal receiving devices and communication equipment on all traffic equipment, and how to unify the signal frequency is also a difficult problem for researchers.
