Nowadays, autonomous driving has become a necessary highlight for all car companies when promoting new cars. Many car companies are trying to publicize how amazing functions their automatic driving systems can achieve, including but not limited to conditional automatic parking, automatic car following to some extent and some auxiliary safety functions. For many people who don't understand cars, these various automatic driving systems seem to be magical.
However, most people who have really used these systems will find that the vast majority of these autopilot systems are still at L2 level, and these L2-level autopilot systems are divided into full-function and non-full-function ones. In addition, L2 itself is not a real autopilot, but only an auxiliary driving system, so most car owners have reservations about these fantastic autopilot systems.
But it is also an automatic driving system. Tesla's can actually realize unmanned driving, unmanned parking, unmanned high-speed cruising, automatic obstacle avoidance in unmanned state and so on. Even Tesla has successfully tested complete unmanned long-distance driving many times. The same is automatic driving. How come Tesla is so good? Where are the gaps?
is it so powerful because the hardware is particularly strong? That's not the case!
anyone who knows about the automatic driving system knows that this kind of thing essentially depends on the computing power of the system hardware. Theoretically, the stronger the hardware performance, the more functions related to autonomous driving can be realized. So, Tesla's self-driving hardware should be powerful in the world?
No, in fact, Tesla has already published all the hardware used in the autopilot system of each car and the number of corresponding radar models. In theory, if you like, you can copy a system exactly the same as Tesla's autopilot system according to the technical documents published by Tesla. Even Tesla has made some patents for autonomous driving public.
Conceptual advantages
In that case, why can't anyone replicate Tesla's success in autonomous driving? The key point is that Tesla's understanding of autonomous driving system is completely different from other manufacturers in the world. Tesla believes that it is not enough to have hardware for autonomous driving. Even the hardware may not have the strongest performance, but the hardware and software must match the car itself. Even if the hardware of different cars is exactly the same, the whole software and adjustment will be very different. In short, Tesla believes that autonomous driving is a system that is integrated with the vehicle itself.
For most other car companies, autonomous driving is only an auxiliary subsystem, which is not necessary or even dispensable. It is isolated from the core system of the vehicle. Since it is only a subsystem, as long as there is powerful hardware and a set of almost available software, the function of automatic driving can be realized theoretically. However, due to the lack of system-level adjustment and software matching, the autopilot systems of most other car companies are still far from Tesla.
This is like Apple and Android in the mobile phone market. Apple relies on all aspects of ios and hardware to get through, even with poor hardware, it can get a much higher experience than Android phones, even if the latter's hardware configuration is much more powerful than itself. Moreover, according to public sources, the core module of Tesla's autonomous driving system has developed from initial external procurement to today's complete self-research. Today's Tesla can already achieve the degree that urban roads are completely unmanned.
how far are we from the benchmark?
to be honest, it's very far away. At present, in the pure tram market, whether it is the independent brand of joint venture car companies or new forces, Weilai Changan and Tucki have done the best in autonomous driving. Weilai's autonomous driving is based on the world's top hardware level and regardless of the cost of stacking. Chang 'an's autopilot was developed as early as 1 years ago, and now it can almost achieve full-function L3. Tucki's autonomous driving is the fastest among the three, but it is far from the completely systematic level of Tesla.
Therefore, in the foreseeable future, Tesla's dominant position in autonomous driving will not be challenged, and the gap will widen. The industry should also continue to exert more efforts in this regard and strive not to let Tesla pull us too far.
This article comes from the author of Chejia, car home, and does not represent car home's standpoint.