原生AI融合网络数字孪生赋能下一代无线网络自治

摘要：面向下一代无线网络的高水平阶自治需求，提出一种融合原生人工智能（AI）与网络数字孪生的一体化架构。该架构涵盖无线网络数字孪生中的数据采集、模型构建与管理等关键环节，以及原生AI驱动的网络性能预测、AI用例自生成与网络策略自定制等核心能力。通过将原生AI深度融入数字孪生系统，构建了“数据‑模型‑决策‑验证”的内生智能闭环，为应对6G网络高复杂度与高动态性环境下的自治挑战，提供了系统化的架构设计与理论支撑。

关键词：6G；数字孪生网络；原生AI

Abstract: To address the demand for high-level autonomy in next-generation wireless networks, an integrated architecture that converges native artificial intelligence (AI) with network digital twin (NDT) is proposed. The framework encompasses key processes within the wireless NDT, including data acquisition, model construction, and management, alongside core native AI-driven capabilities such as network performance prediction, automated AI use case generation, and self-customized network strategy formulation. By deeply embedding native AI into the digital twin system, an endogenous intelligent closed loop of "data-model-decision-verification" is established. This provides a systematic architectural design and theoretical foundation for tackling the autonomy challenges posed by the high complexity and dynamic nature of 6G network environments.

Keywords: 6G; digital twin network; native AI