人工智能辅助的信道估计最新研究进展

An overview of artificial intelligence assisted channel estimation

作为第六代移动通信发展的主流方向,智能通信正在蓬勃发展中,且初步展示了其与传统通信方法相比的优势。人工智能辅助的信道估计作为智能通信的重要组成,在已有的研究成果中展示了其相比传统信道估计算法的优越性,尤其是基于压缩感知技术、超分辨技术、残差学习等开展的信道估计研究均获得了丰硕的成果。针对人工智能辅助的信道估计技术,结合近来学术界最新研究成果,分别从基于深度卷积神经网络、基于深度循环神经网络、基于超分辨技术、基于压缩感知技术4个维度展示了人工智能辅助的信道估计的全貌。最后,对比总结了4类信道估计方法优劣及其未来研究方向,展望了信道估计与深度学习结合的广阔前景。

As the mainstream of the sixth generation mobile communication development,intelligent communication assisted by artificial intelligence technology is vigorously developing,and has initially demonstrated its advantages over traditional communication methods.As an important component of intelligent communication,artificial intelligence assisted channel estimation shows its superiority over traditional channel estimation algorithms in the existing research results,especially those researches based on compressive sensing technology,super resolution technology,residual learning,etc.Aiming at the channel estimation technology assisted by artificial intelligence,combined with the latest research results in the academic field,the whole picture of the channel estimation technology assisted by artificial intelligence from the four dimensions of deep convolution neural network,deep recurrent neural network,super-resolution technology and compression sensing technology were showed.Finally,the advantages and disadvantages of four kinds of channel estimation methods and their future research directions,and the broad prospect of the combination of channel estimation and deep learning were looked forward.