Position based near-field codebook design and beam training for extremely large-scale ARIS

Mounting reconfigurable intelligent surfaces (RISs) on unmanned aerial vehicles (UAVs) inherits the benefits of both technologies, such as flexible deployment, coverage extension, energy saving, as well as interference and blockage effect mitigation. Hence, it is expected that extremely large-scale aerial RIS (XLARIS) will play a significant role in boosting the performance of the sixth-generation communication networks. However, the estimation of the end to end cascaded channel between the transmitter (TX) and the receiver (RX) through the XLARIS is more challenging due to the passive nature of RIS. Furthermore, the beamforming design, in practice, is a complicated process, and thus, beam training (BT) based on a predefined near-field codebook (CB) has been generally adopted. However, an exhaustive search along the large conventional CB, to find the optimal XLARIS configuration, will cause extremely high system overhead and power consumption. In this paper, we design a novel positioning based CB (P-CB) for the near-field of the XLARIS with an affordable CB size. Using this P-CB, a low overhead BT process is developed. The proposed P-CB uses the estimated TX and RX positions as initial points for determining the main reflecting path. Then, depending on the standard deviation of the localization error, sets of sampled points are determined for steering the XLARIS. Simulation results validate the superiority of the proposed scheme compared to other CB based schemes in terms of the lower required overhead to obtain high performance. Moreover, even if the localization accuracy decreases, the proposed P-CB scheme still achieves promising results.