Dual LSTM aided channel estimation for mixed RF-FAS/FSO non-terrestrial cooperative relaying system

This paper proposes a deep learning-aided channel estimation scheme for the radio frequency (RF)-fluid antenna system (FAS)/free space optics (FSO) cooperative relaying systems with high altitude platform station (HAPS). The system model comprises of terrestrial network connected to a satellite through an intermediate HAPS as a relay node. Moreover, the HAPS is facilitated with FAS receiver and can operate in both amplify-and-forward (AF) and decode-and-forward (DF) protocols. It is proposed to estimate the channel state information (CSI) of the RF and optical wireless links using separate long short-term memory (LSTM) deep learning models. Performance of the proposed model is evaluated in terms of outage probability and Ergodic capacity for the considered non terrestrial RF-FAS/FSO relaying network, for both types of cooperative relaying systems.