A briefing session with Yong-Jae Kim (Co-founder & CTO), Nara Won (President Americas), Ho-Joung Kim (Marketing Manager) at Smart Radar System.
Smart Radar System (SRS) journey since inception:
SRS entered the business in October 2017 and the first prototype of 4D imaging radar was launched in 2019. With the strong expertise and market viable product design, the company participated in the CES 2020. The same year in June the company started actual deployment of the radar sensors to its diverse customer base in South Korea. Almost 100,000 units were deployed to the Ministry of Health and Welfare for monitoring elderly patients. Realizing the market demand for advanced radars in healthcare, the company launched an imaging radar for indoor applications in Sept 2021.
Smart Radar System has tapped the opportunity in the automotive industry by unveiling automotive-grade 4D Imaging radar in the 3rd quarter of 2020 and commercialized it in the U.S. market in Jan 2021.
“We are taking immense efforts in continuous development and deployment of radar products that are capable of serving diverse applications to promote and ensure uninterrupted safety.”
Through these years Smart Radar System has been recognized at multiple instances by the South Korean Government and media houses.
Headquartered in Seongnam-si, Gyeonggi-do, South Korea, the company has also expanded its business center in Irvin, California, U.S. A specialized R&D center is located at Future Mobility Technical Centre (FMTC) in SiHeung-si city, South Korea, next to the Seoul National University (SNU) campus.
“The strategic placement of our R&D center at FMTC facilitates us with the SNU’s already built autonomous driving test track helping us to conduct on-road testing of our 4D imaging radar solutions.’’
The company has onboarded a strong R&D enthusiastic team to gain a competitive edge in the radar industry. Nearly 74% of its total resources are working on research and development role to offer technological robustness.
Being a R&D-centric company, SRS has a setup of cabs for testing in-cabin monitoring radars and an autonomous driving track for testing 4D Imaging radars in vehicles. It also has a complete set-up for healthcare application testing such as patient monitoring.
Product Approaches and Business Model:
Off-the-Shelf radar chipset:
Unlike other radar players in the market, Smart Radar System is using off-the-shelf radar chipsets manufactured by leading semiconductor companies such as Texas Instruments, NXP, and others. The reason is to attract the customers of these leading semiconductor players and create an opportunity for business expansion through such partnerships.
Unique Antenna Pattern:
Smart Radar System’s radar products design stands out in the market due to its unique antenna pattern approach. The company for its imaging radar technology uses a non-uniform antenna pattern that allows it to offer higher resolution at a reduced cost. The use of a non-uniform antenna pattern requires a smaller number of virtual antennas, reducing the overall post-processing computer power. This technology is said to generate more demand for autonomous driving applications in foreseeable future.
Signal Processing Algorithm:
Smart Radar System has a patented signal processing algorithm that allows the conversion of a non-uniform antenna pattern to the uniform style of data. This critical algorithm allows the accurate calculation of angles with the use of the non-uniform array. The company has taken huge R&D efforts in the development of this signal processing algorithm that makes it technologically viable and efficient for the majority of applications. The company has a global patent for this technology registered in regions such as the U.S., South Korea, and Europe (Patent-pending in Europe).
Smart Radar System works with two business approaches.
1. The first approach is embedded computing i.e., a fully embedded radar module, wherein the company’s radar module is integrated with the DSP port, and all the deep learning, AI, and machine learning algorithms run in the embedded module, reducing the processing burden of the central computer.
Smart Radar System is doing pretty well in the industrial and healthcare sectors by using this approach.
2. Realizing the transition of the automotive industry towards more software-defined vehicles triggered by Tesla, Smart Radar System is using another approach of isolating radar modules from DSP ports. In this approach, all the AI-based signal processing is conducted in the centralized computer by just collecting the raw data from the sensor module.
Smart Radar System has huge competence over its competitors in this approach as it is using non-uniform antenna patterns instead of a uniform antenna pattern that tends to generate huge raw data.
Radar Product Line and Applications:
Smart Radar System has a broad product portfolio including IRIS and RETINA product lines. The IRIS product line is a typical radar generating 3D point cloud whereas the RETINA is an advanced 4D point cloud radar product line. Both are well-commercialized product lines serving a diverse customer base.
The RETINA 4D imaging radar products range from short-range to long-range designed for applications such as ADAS, autonomous vehicles, autonomous off-highway, equipment safety, smart city, and indoor monitoring. This includes RETINA-4F, RETINA-4SN, and RETINA-XF (X > 4) product segments.
Both the single radar IRIS and 4D imaging radar RETINA offer high resolution, wider FoV, and a longer detection range. Below are the general product specifications for different product segments.
Product Specifications of RETINA Product Line
Above are the specifications of the long-range RETINA-4F.
Product Specifications of IRIS Product Line
RETINA product segment consists of RETINA-4F, short-range radar, and RETINA-4FN, long-range radar. RETINA-4F generates the point cloud data for height and velocity enabling precise detection, tracking, and classification of objects. This RETINA-4F product segment is designed to serve majorly the autonomous driving application. SRS has embedded AI and neural network in the RETINA-4F radar module making it capable of detecting and classifying cyclists, bikes, pedestrians, and other objects in any weather condition. SRS has deployed these radars on heavy construction equipment vehicles, autonomous vessels, forklifts, and others for its couple of South Korean clients.
For in-cabin automotive applications, the company is offering both single chip and 4 chip radars depending on the customer requirement. However, the company believes that single-chip radars are the best solution for the in-cabin application as they are very much capable of counting occupants in the vehicle. Also, single-chip radars are a cost-effective solution as they involve a smaller number of antennas. The only challenge is the precise detection of corner cases; however, the company is making efforts to enhance its performance.
RETINA-4SN generates a heavy point cloud and is capable of recognizing human posture, tracking person count without hijacking their privacy. This product is used for indoor application segments such as trains, malls, restrooms, and hospitals.
RETINA-XF is next-generation radar with (X > 4) chip cascades. The autonomous vehicles industry demands denser and higher resolution point cloud data to ensure safety. Smart Radar System is working towards meeting this demand by developing X chip cascade product line RETINA-XF. The company has done several third-party tests to evaluate the performance of the first version of RETINA-XF and the results are quite satisfying as the product offers Azimuth 1-degree and elevation 2-degree resolution. The public announcement and release of this product is expected by early next year and soon then would be commercialized.
SRS also has deployed 4 chip cascade radars for healthcare monitoring application that includes vital sign detection and posture detection of patients. This application has proved viable during the covid-19 pandemic for contactless monitoring of patients. Radar detection can work perfectly in the absence of light and even in curtain-protected areas. These 4D cascade radar solution is embedded with deep learning and offers a higher level of results.
Another application where Smart Radar System is using its single-chip radar is safety monitoring in public toilets, to detect the occupancy and the time for which the lavatory has been occupied. The company is implementing this solution with a few of its customers in South Korea. This unalike application is aimed at providing help in case of emergencies such as a falling person on the floor and over drug doses in non-monitored areas like restrooms.
Moreover, the Heavy Equipment Industry is where SRS is currently deploying and has commercialized radars in North American and South Korean regions. The company’s IRIS and RETINA radars are being used by its customers for detecting obstacles around the working area and avoiding collisions. For this application, the customer base varies from construction vehicles, off-highway vehicles to garbage trucks. The applications of radars for these customers are not restricted to only object detection but more intelligent services such as remote-controlled operations. At present, two radars are being commonly used for detection; however, it is expected to further increase with the increase in demand for intelligent services.
The company is offering single-chip radar to the oil and chemical companies for level meter application. The field testing for this application is ongoing and the company is expected to soon receive volume orders for the level meter application.
SRS provides lots of radar modules for diverse applications and their sub-segments. The company’s veracity in product offerings supporting diverse applications is one of its major strengths.
“As we are using off-the-shelf radar chipsets and DSP, customizing radar solution for each application is relatively simpler and offers us an upper hand over our competitors.”
SRS has a wide customer base for its product line for construction, healthcare, industrial, and smart city applications, where the products are very well commercialized. However, in the autonomous driving application segment, the company is assessing its product engineering sample with several customers. For this application the sample A is ready, and the company is progressing towards sample B and will be ready by the end of this year. It also expects to commercialize autonomous vehicle grade 4D imaging radar products by early 2023.
Though the company is targeting a wide application base, industrial and healthcare markets are its current cash cows and automotive is an opportunistic market for the company.
Challenges and opportunities:
The imaging radar market holds immense potential especially in the autonomous driving industry as it is expected that in coming years, the SAE level 2 and above vehicles would be equipped with 5 to 8 imaging radars to ensure higher level safety. However, the challenge here is the cost of imaging radar.
Smart Radar System is trying to overcome this challenge by outsourcing the chipsets. It is utilizing X MMIC chipsets along with the DSP board from leading semiconductor companies which help it to reduce the overall manufacturing cost and allows quicker mass production.
“The major challenge that the industry is facing is the higher cost of imaging radar, though it is less expensive compared to other sensing solutions such as LiDAR; however, further reducing the per-unit cost of imaging radar is what we are working towards.”
According to the Smart Radar System, the automotive industry is witnessing a major shift towards software-defined vehicles. This transition will have a larger impact on the imaging radar architecture as well. The company is already ready for this forthcoming disruption by having adopted the approach of isolating the RFIC from the computation module. As OEMs are redefining autonomous vehicle E/E architecture, SRS believes that it is high time for sensor companies to adapt to this change.
In centralized E/E architecture, most of the raw data is fed to the central computer, and in this case, if the number of channels increases, the data rate also increases exponentially, which in turn increases the computing power. To this, the solution is to reduce the amount of data being fed from radar sensors to the central computing unit. And this is possible only with a reduced number of virtual antennas.
“Our patented non-uniform antenna array approach allows the radar signal processing to be done in the vehicle’s centralized processor. Our non-uniform antenna requires lesser MMICs, and virtual antennas compared to companies using uniform antenna patterns.”
Smart Radar System has competence in the non-uniform antenna approach and believes that it is approaching next-generation autonomous vehicle architecture. Hence, the company considers this a challenging opportunity in the autonomous driving industry.
The company also has strong competence in real-time machine learning-based target recognition capability solutions. It is targeting the untapped revenue pockets of the industry by developing its proprietary target recognition engine with raw data from radar sensors. The company sees a huge business opportunity in target recognition solutions based on AI and raw data.
Smart Radar is considered to be one of the leaders in the 4D imaging sector. Still, it is certainly a huge challenge for SRS to survive against many competitors including new players and traditional Tier-1s. But SRS with its strong technical capabilities is open to partnering with these companies to further enhance and empower the imaging radar technology. Also, the company is quite confident that its technology is mass production ready, cost-effective, and highly precise. It has positioned itself quite well in the radar market across the North American and South Asian regions and is ready to fiercely compete in the market.
SRS has a stronghold in the industrial and healthcare segments, while it agrees to have scope for advancement in the field of autonomous driving industry.
Short terms and long-term goals:
SRS is currently focusing on developing AI-based, high resolution, high performance, and cost-effective radar solutions for ADAS applications.
Also, in the near future, the company is aiming to get IPO by initially listing on the KOSDAQ market.
The company’s long-term goal is to invest a huge amount in developing low-cost solutions for autonomous driving vehicles and generate big volume deals. Also, the company aims at tightening relationships with OEMs and Tier-1s to expand its read across the globe.
4D Radar Sensor
Smart Radar Sensor
Get to know the penetration of 4D imaging radar in self-driving industry
Insights on In-cabin and Exterior Sensing Applications of 4D imaging radar in ADAS and Autonomous Vehicles; Emerging 4D imaging players competition assessment