‘Tesla vs the rest’ Who will win the 2021 self-driving Lida War?

Tesla “no need” vs. Weimo and Xiaofeng “essential elements”
High-precision measurement possible, but expensive and requires an integrated design
Pay attention to the high level of domestic lidar HW&SW startups

[디지털투데이 추현우·김현우 기자] There is an old debate in the global autonomous driving industry. It is a debate about which technologies and systems are effectively used by vehicles in recognizing surrounding objects. And in the middle of this debate is the’Lidar’ sensor.

Lida is a compound word of light and radar, and refers to a technology that can sense characteristics such as distance, direction, and speed of an object by shining a laser around it. It is used in various fields such as surveillance camera sensors, delivery robots, drones, screen doors, and road traffic systems, and is a core technology that makes up an autonomous vehicle in particular.

The self-driving technology currently pursuing commercialization is divided into cases that employ LiDAR and cases that do not. Tesla is a company that introduced autonomous vehicles that do not use LiDAR, and most companies such as Google Waymo and Xpeng of China are making autonomous vehicles that actively utilize LiDAR.

In a word, it is’Tesla vs. the rest’.

■Elon Musk “Lida, I don’t use it for free”

In 2019, Tesla unveiled the direction of development of Tesla’s autonomous driving technology through a Tesla Autonomy Day. At that time, LiDAR and high-precision maps (HD maps) showed a big difference from the autonomous driving technology of other companies such as Waymo and Ford.

Tesla’s autonomous driving technology does not use lidar to accurately measure surrounding vehicles and objects. And it doesn’t create HD maps from data received from LiDAR. Tesla believes that the lidar sensor is too expensive and cumbersome, and that HD maps are also low-utility technologies because they cannot respond immediately to changes in real-time road conditions.

Eight built-in cameras, ultrasonic sensors, and radar installed in Tesla electric vehicles replace the vacant seats in the lidar and HD maps. The road environment image data received from 1.2 million Tesla electric vehicles scattered around the world is analyzed using deep learning technology. Data analysis and processing is handled by the Neural Net chip (NPU) built into Tesla’s electric vehicle.

Using this method, Tesla is pursuing the development of autonomous driving technology that can respond more accurately and quickly without a lidar. In fact, Tesla’s autonomous driving data collected from road sites around the world amounts to 3.5 billion kilometers. Compared to the waymo, which uses lidar, stays at the level of 32 million km, it has a clear advantage.

Tesla argues that sensors are not important, but data and software and systems that analyze them are more important.

Not only Tesla, but also Daimler of Germany is conducting research to secure autonomous driving technology only with cameras and millimeter wave radars without using LiDAR. In Korea, self-driving startup Forty-to-Dot is planning to implement autonomous driving technology with only sensor fusion technology of camera and radar without lidar, and implement a pilot service in Seoul this year.

Weimo and Xiaofeng “Rida is essential for autonomous driving”

On the other hand, electric car makers such as Waymo, a subsidiary of Google’s autonomous driving company, and Xiaofeng of China, are actively using LiDAR. Unlike Tesla, LiDAR sees it as an essential element in completing autonomous driving technology in terms of accuracy and efficiency.

The completeness of autonomous driving technology is improved through three types of sensors: a camera that shoots images, a radar that calculates the distance and speed of vehicles in front of and behind a radio wave, and a lidar that can make a high-resolution HD map by measuring surrounding objects more precisely. have.

In November of last year, Xiaofeng, who came to the fore with suspicion of copying Tesla’s self-driving car technology, is adopting a lidar sensor from lidar manufacturer Livox supported by drone maker DJI. It is explained that the level of autonomous driving can be further increased by adding LiDAR to the autonomous driving technology that benchmarks Tesla.

LiDAR manufacturers explained that 10 years ago, the lidar sensor, which was worth tens of millions of won per unit, has also fallen to the current million won, and it is possible to embed it without harming vehicle design through miniaturization.

The latest lidar sensor developed by Reebok has lowered the price to the level of 1.5 million won per unit. It is confident that if it goes into mass production, the unit price can be lowered below 1 million won per unit. Velodyne and Lumina, the two major mountains in the field of lidar sensors, have pledged to release products worth 500,000 won per unit in the second half of this year. Waymo, a leader in the autonomous driving market, is reducing costs by developing LiDAR in-house.

■Domestic lidar companies are on the rise…Securing global competitiveness

As companies open their eyes to self-driving cars, Lida has become a key component of self-driving cars. In addition to automakers such as GM, Ford, Toyota, and Hyundai, as well as ride-sharing services such as Uber and Lift, IT industries such as Google Waymo, Korean Naver, and Chinese Baidu consider LiDAR sensors as an important factor in developing autonomous vehicle technology. Are doing.

Most of the global lidar market share is occupied by foreign companies such as Velodyne and Lumina. However, in Korea, software companies such as Seoulrobotics and Autoa2z, and hardware companies such as SOS Lab and Carnavi are securing global competitiveness.

Seoul Robotics, a lidar software startup, is leading globalization by cooperating with domestic lidar maker SOS Lab, as well as automakers such as Mercedes-Benz, BMW, and Volvo, and overseas lidar makers such as Velodyne and Oster.

Seoul Robotics is developing and exporting self-driving lidar software to nine countries, including Germany, and recently drew attention by working with BMW to develop a’autonomous lidar recognition system’ for three years. If the development is successful, the technology will be applied to the BMW production plant. It is also participating in the development of a lidar solution for smart cities with Qualcomm and Velodyne.

Autonomous A2G joined hands with Kakao Mobility and started providing a demand-responsive paid autonomous driving service in Sejong City on December 18 last year, and it became a hot topic. Autonomous A2G is in charge of autonomous vehicle solutions and operation. The driving vehicle is equipped with four lidar sensors, one each on the front, rear, left and right.

Navicom started as a navigation and black box company, and has entered the lidar industry since 2016. It is actively cooperating with automakers to develop an autonomous driving lidar built-in 16CH with pure domestic technology.

SOS Lab is a startup established in 2016 that produces autonomous vehicles, security, and industrial lidar products. The lidar for self-driving vehicles is currently being developed as a fixed lidar sensor for the driver assistance system (ADAS) of Level 2~3 so that it can be applied to Level 4 in the future.

Jeong Ji-sung, CEO of SOS Lab, said, “With the increasing number of autonomous vehicles equipped with LiDAR in Korea, an environment that enables safe autonomous driving is unfolding.” It is expected to expand”.

Apple’s Lida debate sparked… The conclusion is a paradigm shift

Until the first half of last year, the Lida debate was centered on Tesla. While most autonomous vehicle development companies are adopting LiDAR, only Tesla has unveiled the use of LiDAR. As Tesla first started commercializing its unrivaled driving data and autonomous driving service, it seemed that the LiDAR dance theory was empowering.

But last December, the Lida debate reignited. This is when the news that Apple is developing an Apple car equipped with a lidar is known. In addition, the share price of LiDAR sensor manufacturers surged.

Share prices of two representative companies, Velodyne and Lumina, soared more than 30%. Both companies are listed through the Acquisition Purpose Company (SPAC), and despite their high technological prowess, they have not secured such a profit model.

The reason why the stock price of parts makers skyrockets in response to Apple’s rumored news of its entry into electric vehicles is not simply due to the fact that Apple cars are equipped with LiDAR. Apple’s entry into the self-driving electric vehicle market implies a completely different market environment and potential for development.

It means that the existing automobile industry centered on Germany, the US, Japan, and Korea is transforming into an IT-centered technology industry. If global big-tech companies such as Google, Amazon, and Microsoft, led by Apple, jump into the automotive market, the paradigm of the automotive market will inevitably change.

Rida is not only meaningful as a sensor itself, but also as a keyword that symbolizes changes in the market.

Reporters Chu Hyun-woo and Kim Hyun-woo [email protected]

Copyright © DigitalToday Unauthorized reproduction and redistribution prohibited