mmWave Channel Estimation in UAV Enabled Communication Network

Abstract

Unmanned aerial vehicles (UAVs) are revolutionizing telecommunication networks by offering innovative solutions to bridge coverage gaps and extend connectivity. However, to ensure reliable and efficient communication between UAVs and ground stations, accurate channel estimation becomes a critical factor. Therefore, much research has been presented in the field of millimeter wave channel estimation. Although, it is necessary to investigate the role of UAVs as a relay between terrestrial users and satellite networks. In this regard, an estimating channel method is proposed. First, due to the noise removal ability of the autoencoder network, it is used to reduce the received signal noise. Next, channel coefficients are estimated with the help of a designed CNN network. Finally, the combining matrices used at the receiver are updated to improve the SNR. The simulation results show that the designed neural networks for noise removal and channel estimation improve the accuracy. Also, updating the combining matrices at the receiver shrinks the area to scan and consequently more directed beams can be used that improve the SNR. Using the proposed structure, the overall accuracy can be improved around 10% for different SNRs.