The World War of the hottest lidar

  • Detail

Lidar World War

Abstract: with the continuous efforts of Google, Baidu, general motors, Uber and other major players, L4 driverless vehicles are getting closer and closer to mass production. However, the industry still needs to overcome many obstacles before the final goal is achieved

under the continuous efforts of Google, Baidu, general motors, Uber and other major participants that subject the test piece to axial pressure, the following are the most typical data processing points of L. the level 4 driverless car is getting closer and closer to mass production. However, before the final goal is achieved, the industry still needs to overcome many obstacles

when asked "what is the biggest obstacle to driverless mass production", almost all practitioners will mention "lidar is too expensive". Tesla's two fatal accidents and successive rear end accidents also seem to imply that without lidar, mass production of driverless cars can only be a distant dream

therefore, lidar has become the focus of attention of all parties in the industrial chain, and even the "strategic commanding height" that they compete for. In particular, in the fourth quarter of 2017, the battlefield of lidar has become one of the most intense and eye-catching "sub battlefields" in the driverless war, and rounds of "you sing, I come on stage" have been staged

Li Yifan, CEO of Hesai technology, a domestic lidar manufacturer, once divided the customers in the global lidar industry into "test group" and "vehicle factory group". He found that compared with the customers of the test group who are "less in quantity, anxious" and insensitive to price, the vehicle factory's demand for lidar is "large in quantity, high in standard and less in money", and a few hundred dollars is the upper limit they can accept. Due to the limitations of these conditions, they often can only consider solid-state lidar (or hybrid solid-state lidar) with small volume and lower cost

in order to meet the requirements of vehicle manufacturers, participants in the lidar market are developing lidars with better performance, smaller size and lower cost. In addition, some collapsible intelligent concept companies have made preliminary progress and announced mass production plans

at present, what are the advantages and disadvantages of various technical routes of "low-cost lidar", and which is easier to achieve mass production? Who are the main players in the market and how are they progressing? What new trends will emerge in the development of lidar in the future? In this article, we make a simple comb based on the published data

I. mixed solid state and pure solid state

according to the technical route, lidar can be mainly divided into these three categories: mechanical rotary; Mixed solid; Pure solid

we are most familiar with mechanical rotary lidar, such as velodyne's hdl64 line and hdl32. Because of the difficulty of mass production, it is a general trend for mechanical rotary lidar to be replaced by lower cost hybrid solid-state lidar and solid-state lidar. Therefore, in this paper, we focus on mixed solid and pure solid

1. Hybrid solid-state lidar

hybrid solid-state lidar, which is a name created by velodyne, That is, "half is mechanical rotation, and half is solid". Specifically, there are no visible rotating parts in the appearance of the product, but there are still some mechanical rotating parts inside, but this set of mechanical rotating parts is very small and can be hidden (the driving and detection components are fixed, but the scanning system is mechanical rotation) 。

vlp-32c and vlp-16 of velodyne and the latest vls-128 are typical mixed solids. However, the concept of "mixed solids" is no longer used in velodyne, and the word they use is "puck"

at present, the concept of "mixed solid" used in various reports is ambiguous. Although many manufacturers claim that their products are "mixed solid", in fact, the technical solutions may be very different

for example, the mechanical components of velodyne's puck series rotate 360 degrees, and the visual angle is also 360 degrees; While the lidars of ibeo, Hesai technology and innovusion are also known as "hybrid solid state", their horizontal viewing angles are all between degrees, and it is visible that they are not rotating 360 degrees

according to various published materials, compared with the traditional mechanical rotary and "radical" pure solid-state, the hybrid solid-state lidar taking the compromise route has the lowest difficulty in mass production. In the next few years, most of the laser radars that can be mass produced belong to mixed solid state

2. Solid state lidar:

there are no mechanical rotating parts inside. All the horizontal and vertical viewing angles of laser detection are achieved electronically, which has an included angle, so it can reach a certain field of view without rotation. But also because it can't rotate, the scanning angle of solid-state lidar is limited. To realize omni-directional scanning, it needs to arrange multiple in different directions (at least one in front of and behind, generally four)

however, due to the automation of assembly and debugging, the mass production cost and time can be expected, so it has aroused great interest in the industry. Solid state radar is considered as the main direction of laser radar in the future

there are three kinds of common technology realization methods of solid-state laser radar: MEMS, OPA and flash

in addition to these three mainstream technologies, velodyne, luminar, aeye, Hesai technology and other companies have also developed new solid-state radar technology solutions

second, the main players who have made clear the mass production schedule/plan

according to the statistics of the review of contracted cars, up to now, the main lidar players who have made clear the mass production plan are as follows:

at present, almost all manufacturers who claim to make low-cost lidar in the market take velodyne as a target, and they like to talk about their advantages over velodyne whenever they meet people. But in fact, velodyne is also brave in self revolution. It starts with adjusting the technical route and automation to reduce costs

note 1:

vlp-32c although the number of lines is lower than hdl-64, the detection distance and angular resolution are better than the latter. Since hdl-64 is not velodyne's highest performance mass-produced product, and it is about to be eliminated, startups should not always benchmark hdl-64 with their own products; Users and investors should pay particular attention that they can no longer be fooled by the various "advantages" of their own products given by lidar manufacturers over hdl-64

note 2:

vls-128 can classify targets, but the amount of data it collects is particularly large, which has very high requirements for background computing, and the large amount of computing leads to high requirements for the power consumption performance of the computing platform, which plays an important role in national defense, aerospace, chemical industry, metallurgy, coating, construction and other industries. For general passenger cars and low-speed commercial vehicles (l4/l5), vlp-32c is enough, and vls-128 should not be used very well. Vls-128 line will basically be used in the driverless vehicles of travel companies in the future

according to Weng Wei, marketing director of velodyne Asia, velodye's positioning is not as a lidar manufacturer, but as tier2, it provides T1 with lidar technical solutions and packaged components that meet the requirements of vehicle regulations and ASIC level, and then Tier1 produces lidar according to the requirements of vehicle enterprises. "Our positioning is to become a chip company like Intel and AMD"

luminar's lidar saw the scene outside the car

Copyright © 2011 JIN SHI