AIGC-based Evolvable Digital Twin Networks: A Road to the Intelligent Metaverse

Xuehan Chen,Linfeng Luo,Fengxiao Tang,Ming Zhao,Nei Kato

DOI: https://doi.org/10.1109/mnet.2024.3411008

IF: 10.294

2024-01-01

IEEE Network

Abstract:The future Metaverse aims to provide users with an immersive experience in a persistent online 3-D virtual environment at any given time or location. To achieve this goal, the future wireless communication system should provide high quality communication service with ultra-high capacity and reliability, ultra-low latency, and so on. As one of the key technologies to achieve the Metaverse, Digital Twin Network (DTN) is applied to improve the network performance by dynamically mapping the virtual and real wireless system in real-time throughout its lifecycle, but the construction and evolution of the digital twin (DT) models highly depend on real-time sensing data from the physical world. In addition, the Metaverse is not only a high-precision replica of the real world, but also consists of some exploration things extending from the real world to enrich the users’ experience. Recently, Artificial Intelligence-Generated Content (AIGC) is proposed as an efficient method to allow the automatic generation of content and experiences in virtual worlds using Artificial Intelligence (AI) algorithms creatively, which offers fresh concepts to effectively address the aforementioned challenges, and leads to a more immersive and dynamic Metaverse. To this end, this paper aims to depict a novel path to achieve the intelligent Metaverse. First, the frontier capabilities and key elements of AIGC are presented. Then, we discuss how AIGC drives an evolvable DTN, and how DTN is applied to promote the application of AIGC, thereby facilitating the development of the Metaverse. Finally, a case study is presented to show the efficiency of AIGC in the DTN for the Metaverse, and some future directions are given.