한국통신학회 KICS

Workshop on Network Technologies for

6G – Integrated AI and Communications

by Hyunggon Park, Ewha Womans University, Korea 

  The sixth-generation mobile communication network, commonly known as 6G, is expected to be commercialized in 2030, and active research is underway to achieve peak transmission speeds of 100 Gbit/s to 1 Tbit/s and transmission delays one-tenth the level of fifth-generation (5G) communication. In June 2023, the ITU-R completed its work on the IMT-2030 Framework Recommendation (hereinafter referred to as the 6G Vision), which defines six target usage scenarios and 15 performance indicators (KPIs). Moreover, the 3GPP (3rd Generation Partnership Project) is expected to begin discussions to achieve the performance indicators in 2024 and 2025 and complete initial specifications in 2027 and 2028. In addition to the advancement of the typical performance targets of mobile communication systems, such as maximum transmission rate and latency, the main difference between the 5G and 6G is in the target scenarios in which new functions, such as artificial intelligence (AI) and sensing, are used in close conjunction with communication, and their explicit performance indicators. Among them, the discussion on the joint consideration of AI and communication (hereinafter referred to as Integrated AI and Communication), which is expected to revolutionize the network structure, is still in its early stages; thus, research, development, and standardization are expected to begin in earnest with the completion of the 6G Vision Recommendations. 

  In this workshop, we discussed the development direction of network technologies required for Integrated AI and Communication, one of the most important target scenarios of 6G. First, we presented an overview of the 6G Vision Recommendation of the ITU-R and the evolution of the 3GPP network structure for AI and machine learning (ML). Subsequently, we discussed the advancement of the ongoing Network Data Analytics Function (NWDAF) related to 5G-Advanced network intelligence, which is currently discussed in 3GPP standards. Moreover, we discussed the technologies and trends for applications in 5G-Advanced systems and air interfaces for AI/ML. Finally, we presented the latest AI/ML technologies that can be adopted in 6G, such as offline reinforcement learning, federated learning, split learning, and transfer learning/meta-learning technologies, and their adoption to 6G networks. 

  We believe that this workshop helped the utilization of 5G-Advanced in Korea and pave the path for research and development of core technologies for 6G communication networks.