6G物理层AI赋能技术的思考

Thoughts on AI-Enabled Technologies for 6G Physical Layer

移动通信与AI融合是6G最重要场景和特征之一,6G是运行AI应用的坚实信息底座,AI作为一种赋能技术能够必将激发6G新活力。深度神经网络对数据的非线性映射能力能够对通信链路模块实现联合优化,重构传统物理层设计,提升6G网络的性能和效率。首先介绍了6G的架构与特点,并简单介绍了基于AI的信道估计算法、基于AI的解调算法以及基于AI的智能收发机。然后以基于AI的解调制技术为例,搭建原型样机进行技术验证。最后展望了AI技术对物理层设计的赋能。

The integration of mobile communication and AI is one of the most important scenarios and features of 6G,where 6G is a solid information base for running AI applications and AI is an enabling technology to certainly stimulate the new vitality of 6G.The nonlinear mapping ability of deep neural network can realize joint optimization of communication link modules,reconstruct traditional physical layer design,and improve the performance and efficiency of 6G network.This paper first introduces the architecture and characteristics of 6G and briefly introduces the AI-based channel estimation algorithm,AI-based demodulation algorithm and AI-based intelligent transceiver.Then,taking AI-based demodulation technology as an example,a prototype is built for technical verification.Finally,the prospect of AI technology enabling physical layer design is presented.