Channel Estimation for Full-Duplex RIS-assisted HAPS Backhauling with Graph Attention Networks

In this paper, graph attention network (GAT) is firstly utilized for the\nchannel estimation. In accordance with the 6G expectations, we consider a\nhigh-altitude platform station (HAPS) mounted reconfigurable intelligent\nsurface-assisted two-way communications and obtain a low overhead and a high\nnormalized mean square error performance. The performance of the proposed\nmethod is investigated on the two-way backhauling link over the RIS-integrated\nHAPS. The simulation results denote that the GAT estimator overperforms the\nleast square in full-duplex channel estimation. Contrary to the previously\nintroduced methods, GAT at one of the nodes can separately estimate the\ncascaded channel coefficients. Thus, there is no need to use time-division\nduplex mode during pilot signaling in full-duplex communication. Moreover, it\nis shown that the GAT estimator is robust to hardware imperfections and changes\nin small-scale fading characteristics even if the training data do not include\nall these variations.\n