Channel Power Gain Estimation for Terahertz Vehicle-to-Infrastructure Networks

Abstract

The use of terahertz (THz) frequencies has been recommended to achieve high-speed and ultra-low latency transmissions. Although there exist very large bandwidths in the THz frequency bands, THz channels are significantly dynamic and complicated, which is challenging for channel estimation. To improve the energy efficiency of wireless networks, THz channel power gains need to be precisely evaluated for determining optimal THz transmission frequencies and power control. Therefore, this work presents a novel conditional generative adversarial networks (GAN) based channel power gain estimation solution in the THz vehicle-to-infrastructure (V2I) networks with leaky-wave antennas, where the THz frequency has a big effect on the antenna gain, path loss and atmospheric attenuation. Simulation results confirm that our solution can accurately estimate the channel power gains versus the THz frequencies at a fast speed.