Machine Learning Based Clustering and Modeling for 6G UAV-to-Ground Communication Channels
Z. Zhang, Yu Liu, Cheng‐Xiang Wang, Hengtai Chang, Ji Bian, Jingfan Zhang
Abstract
Towards the sixth-generation (6G) wireless communication, unmanned aerial vehicles (UAVs) have been regarded as an indispensable part due to its flexible deployment, wide coverage, and high mobility. This also creates challenges for channel research. Scatterers are normally present in the structure of clusters during UAV communication, and cluster-based channel modeling is significant. In this paper, the variational BayesianGaussian mixture model (VB-GMM) algorithm is proposed for clustering, which takes into account the time-space properties. Cluster tracking is implemented using the multipath component distance (MCD) algorithm. Intra- and inter-cluster characterization, such as the number of clusters, cluster power distribution, angular/delay offset, and angular/delay spreads, are well studied. Moreover, cluster lifetime and birth-death (B-D) properties are extracted and analyzed. Based on these cluster characteristics acquired by machine learning (ML) method, a novel UAV-toground communication channel model is proposed, and a fourstate Markov chain is also introduced to portray the evolution of clusters. Simulation results match well with channel measurements, which verifies the practicality of the proposed model. This paper can give theoretical and technical support for the design and evaluation of UAV-to-ground communication systems