"The key to success in the AI-native era of 6G lies in rapid market entry through digital twin-based design, emulation, and verification."

Keysight Korea held a press conference on the 10th to introduce next-generation technology innovations centered on 6G and Keysight's role in supporting them.

Lee Seon-woo, President of Keysight Technologies Korea, who was in charge of the presentation, said that the rapid evolution of connectivity and computing is changing the structure of the entire industry, and suggested digital twin-based design, emulation, and verification as key strategies in the process of moving toward 6G.

Lee Seon-woo, President of Keysight Korea, said, “AI is deeply entrenched in our daily lives and industries, and the transition to ‘AI native’ is underway in all fields, including communications, automobiles, defense and space, and semiconductors.”

He goes beyond the 'add-on' approach of adding software functions to existing equipment, and implements AI from the beginning at the semiconductor, equipment, and system levels. He emphasized that it should be designed with utilization in mind.

The CEO diagnosed that while 5G has achieved results in some industrial fields such as smart factories, robotics, and autonomous driving, the general consumer experience has been limited.

Accordingly, he said that 6G requires a shift from a 'technology-first' approach to designing for both demand and commercial viability with a willingness to pay.

The industry sees 2029-2031 as the target period for the initial commercialization of 6G, but more importantly is the step-by-step preparation leading up to that point, or the "Road to 6G."

The keywords for 6G are expanded coverage and intelligence. Network slicing, which redistributes resources based on situations such as large-scale events, disasters, and emergencies, must be combined with real-time AI decision-making.

To this end, virtualization, which involves softwareizing functions except for some base stations, is accelerating, and the trend of combining terrestrial networks and non-terrestrial (satellite) communications (NTN) to implement seamless connectivity is becoming more evident.

Attempts to reduce delay by applying onboard computing and beam steering are also spreading at the satellite level.

With the proliferation of AI data centers, power consumption and heat generation have emerged as bottlenecks in the industry.

As the adoption of highly integrated components such as GPUs, HBMs, and DPUs increases, the dual challenge of simultaneously improving performance and efficiency remains.

CEO Lee Seon-woo emphasized the importance of system-level verification, including power and thermal characteristics, from the design stage, saying, “Performance alone has its limitations.”

In this regard, Yoo Kwang-yeol, Director of Keysight Korea, introduced a method of virtually recreating a wireless environment similar to the real one through urban 3D mapping and ray tracing-based channel modeling, and linking the channel profile obtained here to a channel emulator and base station/terminal emulator to recreate the field in the laboratory.

Director Yoo Gwang-yeol said that they are also operating a loop-type verification system that checks for phenomena such as video disconnection (bad connection) when entering urban shadow areas through preliminary simulations and compares and corrects them with on-site driving data. He explained that compared to European sites, signal level errors can range from several dB to tens of dB, but they are continuously improving accuracy by reflecting actual measurements.

Digital twins extend beyond terrestrial cellular to NTN scenarios such as satellite orbit and inter-satellite handovers.

In the aerospace field, vibration, acoustics, and temperature An application was also presented to examine the reliability of components by virtually reproducing physical environments, such as stress, under harsh conditions.

In Korea, the use of city-level digital twins is increasing, including in government 6G projects, NTN demonstrations, disaster response, underground utility management, and urban wind path prediction.

AI models, for which it is difficult to obtain real-world data, begin training with high-quality synthetic data generated by digital twins, and then refine the model by retraining it with real-world data collected after deployment.

Conversely, AI iterates predictions and optimizations within digital twins, accelerating the exploration of network, data center, and terminal design.

Director Yoo Kwang-yeol stated, “The structure in which virtual and real platforms complement each other and circulates determines both performance and marketability.”

The meeting also mentioned the need for collaboration with domestic AI semiconductor, memory, and packaging companies.

It is pointed out that system-level verification covering all layers, from semiconductors, boards, racks, clusters, and data centers, is required to prevent GPUs, HBMs, NICs, and switches from becoming bottlenecks in the network fabric.

Lee Seon-woo, President of Keysight Korea, emphasized the integrated support system that combines software simulation and measurement/emulation equipment to integrate design, verification, and optimization into a single flow.

The CEO stated, “The conditions for success are ‘fast failure and fast market entry,’” and added, “The team that can first virtually reproduce and resolve on-site issues and then closely follow the cycle of recalibrating them with actual data will be the winners in the 6G era.”