Text/Shore Qing
On the road of intelligent assisted driving, the new forces of car manufacturing can be described as rolling up again, from the initial Xiaopeng P5 to the lidar on the body, becoming the world's first mass production intelligent car equipped with lidar, many new forces can not sit still.
Models equipped with lidar have sprung up, and the number of lidar can be described as more than one, four mecha dragons, 11 three Avita, two Xiaopeng G9, and then the ideal L9, Weima M7, Zhiji L7, Weilai ET7, Xiaopeng P5, Nezha S have also gone to the road of lidar.
There is no doubt that lidar has become a "fragrant food" for domestic car companies, and is a major shortcut to solve the problem of intelligent driving perception. But A honey B arsenic cream, not everyone maintains a keen attitude towards lidar, such as Tesla, which made a big fuss about automatic driving, but voted against lidar.
As early as three years ago, Musk said, "Whoever relies on lidar is doomed to failure", and therefore Tesla only uses a pure visual solution in intelligent driving - camera + algorithm, lidar is really fragrant, but the high cost is destined to be only installed in high-end models, and it cannot become the choice of "technology HP" in a short period of time.
Perhaps because of the high cost, Toyota is also "far away" from lidar, recently, Toyota's Woven Planet Holdings company has adopted a new approach in the development of autonomous driving technology, using low-cost cameras to improve the automatic driving system, rather than using expensive vehicle sensors such as lidar.
It is not difficult to see that after Tesla, Toyota has also embarked on the camera instead of lidar program, according to the interview information, Woven Planet is currently able to use low-cost cameras to collect data, and effectively train the autopilot system, and achieve the same effect as the use of expensive sensors.
In the future, Woven Planet's team will continue to develop in this direction, according to the person in charge, the pure visual perception scheme has made a breakthrough, which may help reduce costs and better expand Toyota's autonomous driving technology.
On the one hand, domestic car companies ushered in the first year of lidar, on the other hand, the low-cost advantage of camera + algorithm was mentioned by more car companies, so in the matter of intelligent driving, who is better than lidar and camera?
In fact, there is no difference between the two, just to say that each has its own advantages and disadvantages, as we all know, the initial assisted driving is to rely solely on cameras as a perceptron, by collecting a large number of, diverse camera data, to improve the automatic driving technology, which undoubtedly requires a lot of time, invest more team size to test.
At this time, with everyone's eagerness to achieve a higher level of automatic driving, lidar appeared in the field of vision of everyone, and became the best alternative to higher precision perception solutions, because the detection range is large enough, much higher than the camera and millimeter wave radar, etc., so the model equipped with lidar can more accurately obtain the three-dimensional information of the target, and the resolution is high, the anti-interference ability is strong, and the detection range is wide. This provides more data for the central system to continuously serve autonomous driving without the need for huge road condition information and test data.
As Tesla AI head Andrej Karpathy puts it, "Lidar is really a shortcut that avoids the basic problem of visual recognition," reducing time and manpower investment in achieving higher levels of intelligent assisted driving, but with it comes high costs.
It is more like the "showmanship" launched by car companies in order to achieve technological breakthroughs, which can quickly achieve and reach the automatic driving level of L2+/L3+, but based on price reasons, it is not suitable for large-scale wide promotion, unless the cost of lidar in the future can be reduced to the same price as the camera, of course, if there is really a day, presumably a higher-spec perception system has long replaced the status of lidar.
Write at the end:
From a longer-term perspective, the current technical route of car lidar may not be the ultimate choice of car companies, but it is the best choice at the moment, and we are also looking forward to the infinite possibilities of technological development, and on the day when the cost of lidar falls in the future, perhaps it will become as ubiquitous and easy to use necessities as cameras.