The digital twin (DT) has recently been forth in the rapid advancements at cloud computing and artificial intelligence (AI). It has numerous applications in smart cities, Industrial 4.0, internet of things (IoT), etc. In the digital space, the DT creates a multiphysics mirror integrated into the physical system. Status information was supplied into the microgrid DT of fast simulations running concurrently with the actual time simulation. Information could be immediately assessed to identify high-risk network segments. The simulator also supports significant information sharing with the higher application domain, especially computing and evaluating, to match the actual work requirements of the real-world situation. We propose a decentralized DT network implemented at the edge device using edge computing technology, where many AI models are hosted to address various application challenges, enabling the autonomous vehicle support system with low delay and a wide range of applications. We outline the fundamental DT communication methods, list the unresolved problems, and demonstrate that DT communication offers a revolutionary method of cutting-edge multiagent system (MAS) by cloud computing fusing wireless communications and AI.