Computer Vision-Based Channel Prediction for Vehicular Communications

Integration of communication and artificial intelli-gence is an important part of future evolution of 6G. As an important research field of artificial intelligence, computer vision also has broad cross-application prospects in wireless channel prediction due to its special advantages such as not occupying spectrum resources. In this paper, we first carry out channel sounding and environment sensing campaigns, and simultaneously acquire channel and image data. Then, we propose a channel characteristics prediction model based on computer vision, which can realize prediction of typical channel characteristics such as path loss, K-factor and root-mean-square delay spread. Finally, based on measured data, we validate the performance of model. The results demonstrate that proposed model can accurately predict the above three characteristic parameters with root mean square error less than 3 dB.