A recent development in the field is the emergence of a novel Internet of Things (IoT) framework known as the Artificial Intelligence of Things (AIoT). Based on this framework, it can enable smart autonomous vehicles, unmanned aerial vehicles, and intelligent industrial production lines to effectively reduce costs and expand the scope of sensor perception by utilizing collaborative sensor networks and shared computing resources provided by edge computing. The sensor network formed by traditional internet has successfully achieved global unified wide area connectivity. However, for a certain control object, sensors that are far away from it do not need to participate in the control process, while local existing sensors fail due to the movement of the control object in a specific direction. So this requires a new type of network that can dynamically update the domain. To enhance the real-time responsiveness of regional communication involving controllers and controlled devices, it is essential for the network to establish dynamic local connections that adapt to the spatial location changes of cars, unmanned aerial vehicles, or production line logistics. As a response to this requirement, this paper introduces a novel edge computing distributed communication network and a dynamically updated domain identity architecture. Every endpoint node has the capability to autonomously form a local cluster based on its spatial location for signal transmission. By dynamically polling and accessing this cluster, it achieves close sharing of sensors and AI computing power, leading to a cost reduction in building a comprehensive intelligent control system compared to traditional single-machine AI systems. This holds significant potential for advancing the implementation of comprehensive AI projects.