LiDAR (laser radar) has taken off after being branded by two industry giants, Huawei and Apple.
Since the beginning of this year, both industry and capital market attention to this technology has been growing. On the one hand, the terminal market responded positively, and a number of car companies declared that the company will have models equipped with lidar; On the other hand, startups continue to emerge in the industrial chain, and even some manufacturers in other fields begin to layout relevant technologies and products.
It is worth thinking about, behind everyone's stampede, the actual demand situation of the market? What problems are the laser radar industry facing in the process of development?
Demand status at the beginning of development
Although the application of lidar in the civil market was relatively rare before, the demand of the civil market began to emerge slowly, benefiting from the upgrading of autonomous driving level and the gradual realization of intelligent driving vehicles. That means the lidar industry can't grow without autonomous driving, which is why the industry has had its ups and downs over the past few years as autonomous driving has evolved and stalled.
"There was a high point in the evolution of autonomous driving in '15 or' 16, when a lot of lidar companies were founded," LuminWave executives recalled in an interview with Jiwei. "However, the commercialization of autonomous vehicles didn't go well at that time, so the popularity of the technology slowed down between 17 and 19."
"Until autonomous driving technology has achieved great development in the past two years, although it has not reached the stage of driverless driving, but more and more cars use functions like assisted driving; This brings a new wind to the lidar industry."
The author entered keywords in Baidu search engine and found that lidar and intelligent driving cars have formed an inseparable relationship at the present stage. Also relying on the above background, the vehicle market has become the preferred target of product promotion of many manufacturers in the lidar industry chain.
Although the concept of autonomous driving has been around for a long time, the development of the technology has been slow due to driving safety concerns and restrictions, and most automakers have taken a more conservative approach to it. Until the emergence of the assisted driving system, it gradually improved the situation.
A senior executive of Luowei Technology said: "At the present stage, assisted driving has become an important way to achieve differentiated competition between car companies, but according to the level of assisted driving, car companies choose different solutions. At L2 level, a lot of manufacturers would actually choose a pure vision solution for cost reasons, but above L3, most manufacturers consider lidar an essential component."
However, the company admitted that despite the new direction of assisted driving and the announcement of a few manufacturers to pre-install lidar products this year, the overall market is still in the early stages of development. At present, the shipping volume of the lidar industry chain is generally small, ranging from thousands to hundreds.
"At this stage, one of the most obvious features is the lack of standards in all aspects. Whether it is the technical solution of the manufacturer, the application scenario of the product, or the demand of the customer, it is constantly changing. At least in our opinion, the current stage is still a long time from the maturity of the industry. "But it's more positive than negative for startups, because once there's a solid standard, there's less room for innovation and less room for competition in the marketplace, and there's more opportunity in the current situation."
Two oriented trends
Among the major factors to promote industrial development, technological progress can promote industrial innovation and development, while cost reduction can promote the growth of industrial scale and penetration rate, which are two important conditions for the lidar industry to achieve rapid scale.
First of all, from the perspective of technology, the industrial environment in the early stage of development leads to different end demands in different fields. "At present, our products are not only used in passenger cars, but also in some logistics vehicle transportation scenarios. By contrast, the two scenarios have completely different requirements for products." Luo Wei technology executives said.
On this basis, based on the silicon photonic chip technology, Luowei Technology plans its products into two categories: "large field of view and short range" and "high performance and long range". The former is mainly the close-range scene within 100 meters, the near-field blind repair of vehicles with large FOV requirements and other non-public road scenes, while the latter is mainly the long-distance scene above 200 meters. This kind of product is usually installed in the forward main lidar of the vehicle for ranging and perception under the high-speed scene."
According to the company, last year has launched a prototype for close-range scenarios, this year gradually began to deliver samples to some customers, so far has received a number of batch orders, is expected to achieve mass production next year. For long range products, the LiDAR chip has completed multi-chip rotation and is currently in the chip testing and verification stage. The earliest next year will be the completion of prototype development and delivery of samples to strategic partners.
When talking about the development direction of the technology, the other party pointed out: "The 360-degree mechanical lidar has been tested on many vehicles from 10 years ago. After verification, people generally believe that it is difficult to pass the vehicle specification, and can not really achieve mass production in the field of vehicle; In the last two years, lidar has evolved to semi-solid or hybrid solid-state solutions. Compared with mechanical solutions, the cost has decreased to a certain extent. Moreover, some mass-production models have adopted or plan to adopt Lidar solutions, but the cost is still too high for most automakers. "The general expectation in the industry is that pure solid-state chip-level solutions are more likely to become mainstream in the next three to five years, when autonomous driving and assisted driving capabilities really explode."
Compared with semi-solid/hybrid solid state solutions, which are more common at present, pure solid-state lidar has more advantages in cost, size and reliability. In addition to high reliability to meet the demands of autonomous driving technology, cost reduction is also an imperative process for the industrialization of Lidar.
Set micro network has learned that at present, when choosing Lidar, domestic and foreign automobile enterprises first pay attention to long-distance products. The corresponding scene is the car driving on the highway with a relatively single road environment, and can realize automatic driving through the long-distance forward main radar. However, the current mechanical and semi-solid solutions to achieve long range applications are generally more expensive, the detection range of more than 200 meters of high-performance lidar unit price can reach thousands of dollars or even thousands of dollars.
"The Lidar technology is really starting to be used in large numbers, and the unit price of the product is going to have to come down to at least a few hundred dollars, and if it's going to penetrate into the mid-to-low price range, it's going to have to come down to under 500 dollars. "We think that's the price that's going to happen in the future, and that's what we've been thinking about from the very beginning."
In fact, since the application of the assisted driving system in the civil vehicle market, it is not uncommon to find solutions that can be realized. In addition to the Lidar, which has attracted high attention recently, the "vision +AI solution" has also won the support of auto companies represented by Tesla.
So, are there applications where lidar can't be used?
In this connection, a senior executive of Luowei Technology told Jiwei Network, "Although there are indeed many scenarios that can be satisfied by the vision +AI solution in theory, some scenarios encountered in the process of human driving are infrequent edge cases, which need to be judged and dealt with through human's past experience and common sense. However, the current deep learning AI algorithm is fundamentally different from the processing model of human brain. It is difficult to add numerous edge cases and common sense into the deep learning model only by using the data of driving scenes. We may be able to make AI models more similar to the human brain through technological evolution in a few years, but we can imagine that the time, effort and expense required to achieve this result is enormous, and most car companies around the world will not choose to wait for this solution to be perfectly implemented."
"From both a technical and a business point of view, vision plus lidar is a good solution for autonomous driving in a variety of scenarios," he said. "Many autonomous driving teams think these kinds of sensing solutions have reached a practical level." From a market perspective, automotive companies can even provide customers with higher safety redundancy through Lidar to meet their differentiated functional needs. Therefore, most automobile enterprises tend to use the multi-sensor fusion method including laser radar to achieve automatic driving, which is the easiest, most secure and safest way to achieve automatic driving at present.
