▲Professor Kim Pil-su, Department of Future Automobile Engineering, Daerim University

“Autonomous Driving Sensor Fusion: Lidar is the Most Important”

Lidar, specialized in long-distance perception…provides a basis for stable driving
Hyundai Motor Group, Weak Autonomous Driving Sensors… Must Use Lidar

As autonomous driving technology has recently become less mature, skepticism is growing.

Many companies, including global manufacturers, are slowing down the development of self-driving cars or reducing their research and development costs.

Even in the case of San Francisco, USA, which is the most representative test bed, GM's Cruise, one of the two companies, has been suspended due to limitations caused by a robotaxi accident, and overall, there is a widespread negative view.

Additionally, Apple Car, which had been planned to be commercialized for the past 10 years, recently announced that it would stop development, which became a global issue.

Apple Car was aiming for Level 4 autonomous driving technology, but it was halted due to the lack of technological maturity compared to the investment.

This withdrawal could be seen as more of a postponement than an abandonment of the Apple Car.

This is because the technologies that have been recruited and researched over the past 10 years are significant and will be commercialized again at any time if the technological maturity increases in the future.

Ironically, looking at the Foxconn electric vehicles already being sold in the Taiwanese market, it is possible to think that Foxconn could be the company in charge of the future Apple Car in terms of connectivity with the Apple Car.

Foxconn, which already produces Apple iPhones, has been focusing on researching electric vehicles for the past eight years and is already producing high-quality mass-produced electric vehicles.

It is also an issue to look forward to in the future.

Conversely, there are increasing cases of more intensive development of autonomous driving technology.

This is because, if applied to various fields, including direct autonomous vehicles, it can take the lead as a global business model in the future.

Xiaomi is shaking up the global market by launching the SU7, a highly competitive model, into mass production for the first time.

This model is gaining attention as a model that aims for the best cost-effectiveness with its low price and quality, while also having some full parking functions and autonomous driving functions.

In particular, Tesla, an icon of electric vehicle innovation, is receiving a lot of attention after Elon Musk declared that he would release a robo-taxi with Level 4 autonomous driving capabilities on August 8.

Of course, with the recent slump in electric vehicle sales, there are suspicions that this robotaxi was introduced as a means to overcome the current difficulties, but it is receiving a lot of attention and interest.

We recently met with the Chinese Prime Minister in a hurry to select the Chinese market as a place where robotaxis can be implemented, and in fact, FSD is applied only to models manufactured at the Shanghai factory. Permission has been granted.

In particular, one of the reasons Tesla has become so popular as an absolute powerhouse in the global market is its autonomous driving technology called FSD.

Currently, to properly utilize this feature, you have to pay a fee of several million won or more.

However, in terms of completeness, it is judged to be at level 2, not level 4.

The reason Hyundai Motor Group also increased capital by over 1.2 trillion won in Motional, a U.S. joint venture that has been specializing in autonomous driving technology, is because it believes that now is the time to strengthen autonomous driving technology.

In order to improve the perfection of autonomous driving technology, deciding which sensors to use is an important starting point.

Typically, a mix of radar sensors, ultrasonic sensors, cameras and, most expensively, lidar sensors are used.

Along with the combination of these sensors, how well the algorithms are integrated to make high-density, high-resolution judgments is also a key technology.

Most autonomous driving development companies are utilizing lidar sensors, but their application has been limited due to their very high price compared to other sensors.

Just 10 years ago, the price of a lidar sensor was tens of millions of won, which was a huge deal, but with continuous development efforts, it is now possible to use it for just a few million won, and the price is expected to drop further in the future.

Tesla still only uses cameras in FSD.

Efforts are being made to increase the level of perfection through the fusion of algorithms centered on seven cameras, but the limitations are increasing.

Some even mention that this is the direct reason why it remains at Level 2.
/> This is because in the early stages of development, it was composed only of practical cameras without using ultra-expensive lidar sensors.

At this point, it is difficult to change the direction of development as the issue has already progressed to utilize the LiDAR sensor whose price has fallen.

Moreover, considering the price limitations, we are sticking with only the camera.

Recently, it has been revealed that Tesla has recognized the importance of lidar sensors for several years and has been focusing on research and development by attaching lidar sensors to its vehicles. This puts Elon Musk, who has always argued that lidar sensors are useless, in an awkward position.

The important part is that the use of lidar sensors in global autonomous driving research is the most important factor in increasing overall density and resolution and in configuring the overall completeness.

Although each sensor has its own strengths and weaknesses, the advantages of the lidar sensor are more prominent in that it provides a basis for covering 360 degrees, reliably recognizing traffic lights and long distances, and moving safely.

Even if you wanted to use a lidar sensor, there was a lot of pressure to use it because Germany's Valeo was the only one, but recently, domestic startups have succeeded in developing lidar, expanding the options.

In addition, expectations are growing as Hyundai Motor Group has recently made strong investments to develop autonomous driving technology.

What is clear is that Hyundai Motor Group is at a lower level than other advanced companies in terms of autonomous driving sensor and algorithm development.

First of all, you need to use a lidar sensor that is competitively priced to maximize all functions.

The advantage of lidar sensors is that they are very reliable compared to other sensors. I certainly hope this will be a solid starting point for autonomous vehicles from the get-go.

Now is the most important time.