Accelerating multisensor fusion and accelerating the evolution of single line lidar on the road is imminent.
As the core visual component of intelligent driving or automatic driving technology, this year's lidar market is very busy. The continuous mergers and acquisitions by the giants, the successive release of various types of products on the market and the growing number of auto drivers in the LiDar camp have emerged, and laser radar has become a major symbol of the era of automotive intelligence. However, in view of the current market situation, the speed of LiDar is still very slow. Therefore, more and more manufacturers began to push the cost of single line lidar products, so as to "pioneer" to knock on the door of the automotive application market. Shenzhen city Leishen Intelligent System Co. Ltd. CEO Hu Xiaobo
"Before the cost of multi line laser radar has not been reduced, I think the single line laser radar can be used as much as possible before the cost of the multi line laser radar has not been reduced," CEO Hu Xiaobo, a Shenzhen laser Intelligent System Co., Ltd., stressed. In fact, the single line in the front and back of the anti-collision, around the blind area detection and automatic parking and other aspects can play a very good effect, so that the cost can be further low. There was a statistic before, saying that most of the traffic accidents are mostly side collisions. In particular, more than 40% of these collisions, such as large trucks and buses, account for about 8% in forward collisions, more than 70% in backward collisions, and more than 20% are caused by the blind area detection. In fact, if low cost single line lidar is used up, it can solve this problem well, and it should reduce more than 80% traffic accidents every year. There are also some car factories testing our single line lidar. It is still good from the test results. In the future we want to use lower costs and prices to make the single line laser radar use as fast as possible, like this single line of 100 to 150 meters, and then use it in the car for the future. From 1000 yuan to 1500 yuan, it will make all cars affordable.
As Hu Xiaobo said, today's automobile sensor portfolio is far from the "safety" requirements, especially the huge trucks and trucks. Although some vehicles are equipped with other sensors such as millimeter wave, camera and ultrasonic wave, there are many accidents caused by the blind area detection and the difficult location of the vehicle. Therefore, the "ahead of the car" of the laser radar is imminent. However, future intelligent driving is still dependent on multi sensor fusion technology. After all, from the perspective of cost performance, reliability, security and stability, a single laser radar or other sensors can not guarantee full security redundancy.
Hu Xiaobo agreed, he said: "laser radar, camera and millimeter wave sensors are not in fact who replace who, I think each other is a complementary problem." Because the car is a very high demand for security, if you say which sensor is out of the problem, the other is not redundant, it will be very dangerous. For multi-sensor, laser radar is good at obstacle detection and obstacle tracking, and some details, such as lane lines, are not measured, and the camera is good at distinguishing the color information of the environment, and can well identify the lane line, the pedestrian, the car and the surrounding environment, millimeter wave mine. Da can quickly get speed information, and it has low attenuation and penetration in fog. There is a complementary relationship between the three.
At present, the laser also developed a multi sensor fusion ADAS system based on single line laser radar and camera. Hu Xiaobo added: "in this, laser radar is used to measure point cloud data, and then combined with camera data to do lane line recognition, lane departure warning and curvature data, so that the distance of the car can be very long. Accurate measurement. This data is relatively simple, but at present a lot of car factories are generally difficult to test the laser radar point cloud data, and through our processing, we can greatly simplify the data. After these data processing, can be sent to the central controller through the CAN bus, so that the car factory can make a good decision.
In the future, with the continuous development of intelligent driving technology, the editor believes that the solution of multi sensor fusion based on laser radar + high precision map + cloud computing will continue to dominate. Under the strong impetus of the key "infrastructure" of the laser radar, the future vehicle autonomous environment perception, positioning, high precision 3D map data collection and drawing will also be given powerful hardware kinetic energy, thus promoting the automobile industry towards a highly intelligent age.