[TECH Weekly] The Autonomous Driving Rescuer Breaking Through Heavy Rain and Fog: Smart Radar System Leads 4D Imaging Radar Innovation

Autonomous vehicles face visual limitations in bad weather and at night when relying only on optical cameras and LiDAR. Optical cameras are vulnerable to lens contamination and blocked visibility, while LiDAR is vulnerable to high costs and physical impact, creating environmental constraints that make it difficult for autonomous driving to respond with cameras and LiDAR alone.

A next-generation sensor that can overcome these limitations is emerging. The 4D imaging radar, regarded as a key technology for future mobility, can precisely identify the distance, height, and speed of road objects even in harsh environments, and this identification data is three-dimensional spatial data. The importance of technology that can grasp object location, shape, and attributes in three dimensions, even under adverse conditions, is growing.

Smart Radar System is presented as a representative case of a deep-tech company with unrivaled core technology in 4D imaging radar. The company is pushing ahead with localization of radar technology based on its proprietary deep-tech capabilities and aims to leap forward as a global sensor company.

Smart Radar System, founded in 2017, is a next-generation radar specialist company established by key researchers from KAIST and Silicon Valley in the United States. Since its founding, the company has focused on developing 4D imaging radar technology, which consists of an irregular antenna array and AI algorithms. While conventional 2D/3D radar has been limited to distance and direction measurement, 4D imaging radar generates high-resolution point cloud data including height information.

From the beginning, Smart Radar System avoided components for home appliances and smartphones and instead targeted automotive, robotics, drones, defense, and special mobility as its initial markets. This was a strategy aimed at sectors with a strong need to operate in harsh environments, and the company set its sights on the high-end automotive electronics sensor market.

Smart Radar System was listed on the KOSDAQ market in 2023, and after the listing it consecutively secured partnerships with global automotive Tier 1 parts suppliers and also consecutively secured partnerships with global robotaxi companies. Based on this, it has now established itself as a key partner in the global autonomous driving sensor ecosystem.

Conventional radar had limitations in low resolution, and ordinary 3D radar could not measure object height. Because of this, errors occurred in which obstacles on the road and overhead structures were not distinguished when passing tunnel entrances or under overpasses. Smart Radar System implemented 4D imaging radar based on its core technologies of irregular antenna design and AI signal processing algorithms, thereby extracting the four-dimensional information of distance, height, depth, and speed in real time via radio waves.

Smart Radar System's 4D imaging radar generates LiDAR-level high-density pointing data and maintains 100% performance in heavy rain, fog, and darkness. As a result, it can support the autonomous driving ECU in flawlessly distinguishing every object on the driving path.

Competitors rely on deploying large numbers of expensive semiconductor chips (MMIC) to improve resolution, but Smart Radar System achieved this through the development of irregular antenna placement and AI signal-processing restoration software correction technology. This approach dramatically lowers unnecessary hardware costs while significantly reducing hardware expenses and power consumption. It also has secured a unique mechanism that processes high-resolution 3D spatial data at dozens of frames per second without data processing Latency.

Smart Radar System is carrying out 4D imaging radar supply and joint development projects for global automakers, and it is also carrying out 4D imaging radar supply and joint development projects for Tier 1 partners and North American robotaxi companies. The company is seeking to prove its technological capabilities through these projects.

The range of 4D imaging radar applications is expanding to autonomous passenger cars, commercial vehicles, construction equipment, agricultural machinery, drones, and defense. The application areas are expanding rapidly.

As the commercialization of autonomous driving at Level 3 and above accelerates, mass-production vehicle adoption of Smart Radar System sensors is becoming full-scale, and mass-production vehicle adoption of Smart Radar System software algorithms is also becoming full-scale. Accordingly, Smart Radar System is entering a full-scale mass-production phase, and since its existing revenue structure was centered on development services and sample supply, analysts are pointing to the potential for cumulative high-margin module revenue per vehicle and cumulative software license revenue per vehicle, raising the possibility that the company has reached an inflection point in performance.

The automotive electronics parts industry typically takes several years from product development to silicon and automotive validation to automaker adoption, and it has strong barriers to entry. Smart Radar System has built an order pipeline with global customers over the past few years, and that pipeline is now in the stage of moving toward full-scale series production.

When a standard sensor for automotive electronics is adopted, stable revenue can be generated throughout the vehicle's lifecycle. This goes beyond simple hardware sales and can lead to the status of a high-value-added AI sensor platform company. In that case, profitability could also expand exponentially.

However, Smart Radar System's first practical limitation is price competition with the Vision-Only autonomous driving camp centered on cameras. Its second practical limitation is cost reduction in mass-production processes for passenger cars. Some automakers, such as Tesla, continue to stick with a low-cost camera plus AI software combination instead of high-performance sensors, meaning the relatively expensive 4D imaging radar needs a significant increase in adoption rates, and pressure to lower the price of 4D imaging radar is also growing.

The current supply structure is focused on special mobility sectors such as construction equipment, agricultural machinery, and drones, with low-volume, high-mix production. Accordingly, the transition challenge is shifting the company's DNA to a passenger-car mass-production line at a scale of several million units per year. Key tasks in this process are said to be yield control and material cost control.

In addition, the need to reduce dependence on external semiconductor chipsets (MMIC), lower the module unit price, and maintain high profitability in software licensing has also been raised. The ability to reduce reliance on external chips, cut module prices, and protect software profitability is expected to be the key to a long-term revaluation of the company's value.

A mobility-sector expert said that for pure camera-based systems, as the limits of bad weather become clearer, the installation of 4D imaging radar is inevitable. However, since automakers are demanding lower prices, the key to offsetting that pressure is to add more value to the software algorithms, the expert said. The expert added that whether Smart Radar System, which has accumulated demonstration data in agricultural machinery and special vehicles, can use that data to accelerate entry into passenger-car mass-production lines will be the factor that determines future valuation.

Source: TECHWORLD · Kim Yong-su
Original: https://www.epnc.co.kr/news/articleView.html?idxno=406248

References

• https://www.epnc.co.kr/news/articleView.html?idxno=406248