CogNet: Cognitive Super Resolution Network for Persistent End-to-End Mobility Communication in MIMO Systems

Abstract

The importance of a cyber-edge and cognitive artificial intelligence (AI)-based security strategy to boost autonomous underwater navigation and aerial mobility and prevent hetero-geneous reprisal attacks cannot be overemphasized. This paper proposes a split of super-resolution (SR) to reconstruct the channel state information (CSI) through self-supervised learning for a multiple-input-multiple-output (MIMO) system. Unlike existing designs, this study used a split of SR into two disjoint sub-blocks through transfer learning to improve the CSI detailed structures in the reconstruction process. The simulation results show that the proposed system significantly improved the quality of the CSI after reconstruction compared to the existing system in terms of cosine similarity $\rho$ of 95.2% and normalized mean square error (NMSE) of −16.33 at different compression rates for both indoor and outdoor environments, which is essential for a MIMO system in improving performance, coverage, reliability, and user experience in 5G and 6G networks.